From 0fe3954c06dd6c658f99fff58fc027b720676bf6 Mon Sep 17 00:00:00 2001 From: Aiden Cline Date: Fri, 9 Oct 2026 11:17:23 -0500 Subject: [PATCH 1/4] feat(novita-ai): refresh and audit model catalog --- models/alibaba/qwen-mt-plus.toml | 22 +++++++++++++ models/alibaba/qwen3-embedding-8b.toml | 26 ++++++++++++++++ .../alibaba/qwen3-omni-30b-a3b-thinking.toml | 25 +++++++++++++++ models/alibaba/qwen3-reranker-8b.toml | 26 ++++++++++++++++ models/alibaba/qwen3-vl-30b-a3b-instruct.toml | 25 +++++++++++++++ models/apodex/apodex-1.1-mini.toml | 25 +++++++++++++++ models/baai/bge-m3.toml | 26 ++++++++++++++++ models/baai/bge-reranker-v2-m3.toml | 25 +++++++++++++++ models/baichuan/baichuan-m2-32b.toml | 24 ++++++++++++++ models/baidu/cobuddy.toml | 19 ++++++++++++ models/baidu/ernie-4.5-21b-a3b.toml | 25 +++++++++++++++ .../deepseek/deepseek-r1-0528-qwen3-8b.toml | 24 ++++++++++++++ models/deepseek/deepseek-r1-0528.toml | 25 +++++++++++++++ models/gryphe/mythomax-l2-13b.toml | 22 +++++++++++++ models/inclusionai/ling-3.0-flash-fin.toml | 10 ++++-- models/inclusionai/ling-3.0-flash-sante.toml | 21 +++++++++++++ models/inclusionai/ling-3.0-flash-vl.toml | 25 +++++++++++++++ models/inclusionai/ling-3.0-flash.toml | 25 +++++++++++++++ models/meta/llama-3.2-1b-instruct.toml | 24 ++++++++++++++ models/microsoft/wizardlm-2-8x22b.toml | 24 ++++++++++++++ models/mindlab/macaron-v1-tall.toml | 25 +++++++++++++++ models/mindlab/macaron-v1-venti.toml | 25 +++++++++++++++ models/minimax/minimax-m1-80k.toml | 25 +++++++++++++++ models/moonshotai/kimi-k2-0905.toml | 24 ++++++++++++++ models/moonshotai/kimi-k2-instruct.toml | 24 ++++++++++++++ .../nousresearch/hermes-2-pro-llama-3-8b.toml | 24 ++++++++++++++ models/paddlepaddle/paddleocr-vl.toml | 23 ++++++++++++++ models/sao10k/l3-70b-euryale-v2.1.toml | 24 ++++++++++++++ models/sao10k/l3-8b-stheno-v3.2.toml | 24 ++++++++++++++ .../autoglm-phone-9b-multilingual.toml | 24 ++++++++++++++ .../models/Sao10K/L3-8B-Stheno-v3.2.toml | 10 ++++++ .../models/apodex/apodex-1.1-mini.toml | 13 ++++++++ providers/novita-ai/models/baai/bge-m3.toml | 7 +++++ .../models/baai/bge-reranker-v2-m3.toml | 10 ++++++ .../models/baichuan/baichuan-m2-32b.toml | 22 ++----------- providers/novita-ai/models/baidu/cobuddy.toml | 12 +++++++ .../baidu/ernie-4.5-21B-a3b-thinking.toml | 24 -------------- .../models/baidu/ernie-4.5-21B-a3b.toml | 18 ++--------- .../baidu/ernie-4.5-300b-a47b-paddle.toml | 22 ------------- .../baidu/ernie-4.5-vl-28b-a3b-thinking.toml | 23 -------------- .../models/baidu/ernie-4.5-vl-28b-a3b.toml | 22 ------------- .../models/baidu/ernie-4.5-vl-424b-a47b.toml | 22 ------------- .../models/deepseek/deepseek-ocr-2.toml | 18 ++--------- .../models/deepseek/deepseek-ocr.toml | 22 ------------- .../deepseek/deepseek-prover-v2-671b.toml | 21 ------------- .../deepseek/deepseek-r1-0528-qwen3-8b.toml | 17 ++-------- .../models/deepseek/deepseek-r1-0528.toml | 22 ++----------- .../deepseek-r1-distill-llama-70b.toml | 24 -------------- .../deepseek-r1-distill-qwen-14b.toml | 22 ------------- .../deepseek-r1-distill-qwen-32b.toml | 22 ------------- .../models/deepseek/deepseek-r1-turbo.toml | 18 +++-------- .../models/deepseek/deepseek-v3-0324.toml | 25 --------------- .../models/deepseek/deepseek-v3-turbo.toml | 21 ------------- .../deepseek/deepseek-v3.1-terminus.toml | 27 ---------------- .../models/deepseek/deepseek-v3.1.toml | 27 ---------------- .../models/deepseek/deepseek-v3.2-exp.toml | 25 --------------- .../models/deepseek/deepseek-v3.2.toml | 28 ----------------- .../deepseek/deepseek-v4-flash-0731.toml | 17 ++++++++++ .../models/deepseek/deepseek-v4-flash.toml | 6 +++- .../models/deepseek/deepseek-v4-pro-0813.toml | 17 ++++++++++ .../models/deepseek/deepseek-v4-pro.toml | 5 ++- .../models/deepseek/deepseek-v4.1-flash.toml | 17 ++++++++++ .../models/deepseek/deepseek_v3.toml | 11 +++++++ .../models/google/gemma-3-12b-it.toml | 17 ++-------- .../models/google/gemma-3-27b-it.toml | 16 ++-------- .../models/google/gemma-4-26b-a4b-it.toml | 22 ++++--------- .../models/google/gemma-4-31b-it.toml | 22 ++++--------- .../models/gryphe/mythomax-l2-13b.toml | 19 ++---------- .../models/inclusionai/ling-2.6-1t.toml | 24 -------------- .../models/inclusionai/ling-2.6-flash.toml | 24 -------------- .../inclusionai/ling-3.0-flash-fin.toml | 12 +++++++ .../inclusionai/ling-3.0-flash-sante.toml | 12 +++++++ .../models/inclusionai/ling-3.0-flash-vl.toml | 12 +++++++ .../models/inclusionai/ling-3.0-flash.toml | 12 +++++++ .../models/inclusionai/ling-3.1-flash.toml | 12 +++++++ .../models/inclusionai/ring-2.6-1t.toml | 25 --------------- .../models/kwaipilot/kat-coder-pro.toml | 23 -------------- .../meta-llama/llama-3-70b-instruct.toml | 23 -------------- .../meta-llama/llama-3-8b-instruct.toml | 22 ------------- .../meta-llama/llama-3.1-8b-instruct.toml | 17 ++-------- .../meta-llama/llama-3.2-1b-instruct.toml | 11 +++++++ .../meta-llama/llama-3.2-3b-instruct.toml | 22 ------------- .../meta-llama/llama-3.3-70b-instruct.toml | 22 +++---------- ...lama-4-maverick-17b-128e-instruct-fp8.toml | 16 ++-------- .../llama-4-scout-17b-16e-instruct.toml | 15 ++------- .../models/microsoft/wizardlm-2-8x22b.toml | 17 ++-------- .../models/mindai/macaron-v1-tall.toml | 14 +++++++++ .../models/mindai/macaron-v1-venti.toml | 13 ++++++++ .../models/minimax/minimax-m2.1.toml | 27 ---------------- .../minimax/minimax-m2.5-highspeed.toml | 28 ----------------- .../models/minimax/minimax-m2.5.toml | 28 +++++------------ .../minimax/minimax-m2.7-highspeed.toml | 6 ++++ .../models/minimax/minimax-m2.7.toml | 29 +++++------------ .../novita-ai/models/minimax/minimax-m2.toml | 27 ---------------- .../novita-ai/models/minimax/minimax-m3.toml | 24 ++++++++++++++ .../models/minimaxai/minimax-m1-80k.toml | 19 +++--------- .../models/mistralai/mistral-nemo.toml | 23 -------------- .../models/moonshotai/kimi-k2-0905.toml | 21 ++----------- .../models/moonshotai/kimi-k2-instruct.toml | 18 ++--------- .../models/moonshotai/kimi-k2-thinking.toml | 26 +++++----------- .../models/moonshotai/kimi-k2.5.toml | 31 +++++-------------- .../models/moonshotai/kimi-k2.6.toml | 2 ++ .../models/moonshotai/kimi-k2.7-code.toml | 16 ++++------ .../novita-ai/models/moonshotai/kimi-k3.toml | 19 +++++------- .../nousresearch/hermes-2-pro-llama-3-8b.toml | 20 ++---------- .../nvidia/nemotron-3-nano-30b-a3b.toml | 15 +++++++++ .../novita-ai/models/openai/gpt-oss-120b.toml | 25 +++++---------- .../novita-ai/models/openai/gpt-oss-20b.toml | 23 -------------- .../models/paddlepaddle/paddleocr-vl.toml | 16 ++-------- .../models/qwen/qwen-2.5-72b-instruct.toml | 24 -------------- .../novita-ai/models/qwen/qwen-mt-plus.toml | 19 ++---------- .../models/qwen/qwen2.5-7b-instruct.toml | 22 ------------- .../models/qwen/qwen2.5-vl-72b-instruct.toml | 22 ------------- .../models/qwen/qwen3-235b-a22b-fp8.toml | 22 ------------- .../qwen/qwen3-235b-a22b-instruct-2507.toml | 24 -------------- .../qwen/qwen3-235b-a22b-thinking-2507.toml | 24 -------------- .../models/qwen/qwen3-30b-a3b-fp8.toml | 22 ------------- .../novita-ai/models/qwen/qwen3-32b-fp8.toml | 22 ------------- .../novita-ai/models/qwen/qwen3-4b-fp8.toml | 22 ------------- .../novita-ai/models/qwen/qwen3-8b-fp8.toml | 22 ------------- .../qwen/qwen3-coder-30b-a3b-instruct.toml | 22 ------------- .../qwen/qwen3-coder-480b-a35b-instruct.toml | 24 -------------- .../models/qwen/qwen3-coder-next.toml | 23 -------------- .../models/qwen/qwen3-embedding-8b.toml | 7 +++++ .../novita-ai/models/qwen/qwen3-max.toml | 24 -------------- .../qwen/qwen3-next-80b-a3b-instruct.toml | 22 ------------- .../qwen/qwen3-next-80b-a3b-thinking.toml | 23 -------------- .../qwen/qwen3-omni-30b-a3b-instruct.toml | 25 --------------- .../qwen/qwen3-omni-30b-a3b-thinking.toml | 27 +++++----------- .../models/qwen/qwen3-reranker-8b.toml | 8 +++++ .../qwen/qwen3-vl-235b-a22b-instruct.toml | 22 ------------- .../qwen/qwen3-vl-235b-a22b-thinking.toml | 22 ------------- .../qwen/qwen3-vl-30b-a3b-instruct.toml | 16 ++-------- .../qwen/qwen3-vl-30b-a3b-thinking.toml | 22 ------------- .../models/qwen/qwen3-vl-8b-instruct.toml | 22 ------------- .../models/qwen/qwen3.5-122b-a10b.toml | 25 +++++---------- .../novita-ai/models/qwen/qwen3.5-27b.toml | 25 +++++---------- .../models/qwen/qwen3.5-35b-a3b.toml | 27 ++++++---------- .../models/qwen/qwen3.5-397b-a17b.toml | 25 +++++---------- .../novita-ai/models/qwen/qwen3.6-27b.toml | 15 +++++++++ .../models/qwen/qwen3.6-35b-a3b.toml | 15 +++++++++ .../novita-ai/models/qwen/qwen3.7-max.toml | 29 ++++++----------- .../models/qwen/qwen3.8-2.4t-a95b.toml | 16 ++++++++++ .../novita-ai/models/qwen/qwen3.8-27b.toml | 18 +++++++++++ .../novita-ai/models/qwen/qwen3.8-flash.toml | 14 +++++++++ .../novita-ai/models/qwen/qwen3.8-max.toml | 18 +++++++++++ .../models/sao10K/L3-8B-stheno-v3.2.toml | 22 ------------- .../models/sao10K/l3-70b-euryale-v2.1.toml | 21 ------------- .../models/sao10K/l3-8b-lunaris.toml | 22 ------------- .../models/sao10K/l31-70b-euryale-v2.2.toml | 21 ------------- .../models/sao10k/l3-70b-euryale-v2.1.toml | 8 +++++ .../models/stepfun/step-3.7-flash.toml | 16 ++++++++++ providers/novita-ai/models/tencent/hy3.toml | 17 ++++++++++ .../novita-ai/models/tencent/hy4-preview.toml | 16 ++++++++++ .../models/xiaomimimo/mimo-v2-flash.toml | 26 ---------------- .../models/xiaomimimo/mimo-v2-pro.toml | 18 ----------- .../models/xiaomimimo/mimo-v2.5-pro.toml | 12 +++---- .../models/xiaomimimo/mimo-v2.5.toml | 16 ++++++++++ .../models/xiaomimimo/mimo-v2.6-flash.toml | 13 ++++++++ .../models/xiaomimimo/mimo-v2.6-pro.toml | 13 ++++++++ .../autoglm-phone-9b-multilingual.toml | 19 ++---------- .../novita-ai/models/zai-org/glm-4.5-air.toml | 25 ++++----------- .../novita-ai/models/zai-org/glm-4.5.toml | 29 ----------------- .../novita-ai/models/zai-org/glm-4.5v.toml | 25 +++++---------- .../novita-ai/models/zai-org/glm-4.6.toml | 29 +++++------------ .../novita-ai/models/zai-org/glm-4.6v.toml | 24 +++++--------- .../models/zai-org/glm-4.7-flash.toml | 25 +++++---------- .../novita-ai/models/zai-org/glm-4.7.toml | 29 +++++------------ .../novita-ai/models/zai-org/glm-5.1.toml | 25 ++++----------- .../novita-ai/models/zai-org/glm-5.2.toml | 4 +-- .../models/zai-org/glm-5.3-flash.toml | 18 +++++++++++ .../novita-ai/models/zai-org/glm-5.3.toml | 15 +++++++++ providers/novita-ai/models/zai-org/glm-5.toml | 28 +++++------------ providers/novita-ai/provider.toml | 12 +++---- 174 files changed, 1455 insertions(+), 2088 deletions(-) create mode 100644 models/alibaba/qwen-mt-plus.toml create mode 100644 models/alibaba/qwen3-embedding-8b.toml create mode 100644 models/alibaba/qwen3-omni-30b-a3b-thinking.toml create mode 100644 models/alibaba/qwen3-reranker-8b.toml create mode 100644 models/alibaba/qwen3-vl-30b-a3b-instruct.toml create mode 100644 models/apodex/apodex-1.1-mini.toml create mode 100644 models/baai/bge-m3.toml create mode 100644 models/baai/bge-reranker-v2-m3.toml create mode 100644 models/baichuan/baichuan-m2-32b.toml create mode 100644 models/baidu/cobuddy.toml create mode 100644 models/baidu/ernie-4.5-21b-a3b.toml create mode 100644 models/deepseek/deepseek-r1-0528-qwen3-8b.toml create mode 100644 models/deepseek/deepseek-r1-0528.toml create mode 100644 models/gryphe/mythomax-l2-13b.toml create mode 100644 models/inclusionai/ling-3.0-flash-sante.toml create mode 100644 models/inclusionai/ling-3.0-flash-vl.toml create mode 100644 models/inclusionai/ling-3.0-flash.toml create mode 100644 models/meta/llama-3.2-1b-instruct.toml create mode 100644 models/microsoft/wizardlm-2-8x22b.toml create mode 100644 models/mindlab/macaron-v1-tall.toml create mode 100644 models/mindlab/macaron-v1-venti.toml create mode 100644 models/minimax/minimax-m1-80k.toml create mode 100644 models/moonshotai/kimi-k2-0905.toml create mode 100644 models/moonshotai/kimi-k2-instruct.toml create mode 100644 models/nousresearch/hermes-2-pro-llama-3-8b.toml create mode 100644 models/paddlepaddle/paddleocr-vl.toml create mode 100644 models/sao10k/l3-70b-euryale-v2.1.toml create mode 100644 models/sao10k/l3-8b-stheno-v3.2.toml create mode 100644 models/zhipuai/autoglm-phone-9b-multilingual.toml create mode 100644 providers/novita-ai/models/Sao10K/L3-8B-Stheno-v3.2.toml create mode 100644 providers/novita-ai/models/apodex/apodex-1.1-mini.toml create mode 100644 providers/novita-ai/models/baai/bge-m3.toml create mode 100644 providers/novita-ai/models/baai/bge-reranker-v2-m3.toml create mode 100644 providers/novita-ai/models/baidu/cobuddy.toml delete mode 100644 providers/novita-ai/models/baidu/ernie-4.5-21B-a3b-thinking.toml delete mode 100644 providers/novita-ai/models/baidu/ernie-4.5-300b-a47b-paddle.toml delete mode 100644 providers/novita-ai/models/baidu/ernie-4.5-vl-28b-a3b-thinking.toml delete mode 100644 providers/novita-ai/models/baidu/ernie-4.5-vl-28b-a3b.toml delete mode 100644 providers/novita-ai/models/baidu/ernie-4.5-vl-424b-a47b.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-ocr.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-prover-v2-671b.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-r1-distill-llama-70b.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-r1-distill-qwen-14b.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-r1-distill-qwen-32b.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-v3-0324.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-v3-turbo.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-v3.1-terminus.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-v3.1.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-v3.2-exp.toml delete mode 100644 providers/novita-ai/models/deepseek/deepseek-v3.2.toml create mode 100644 providers/novita-ai/models/deepseek/deepseek-v4-flash-0731.toml create mode 100644 providers/novita-ai/models/deepseek/deepseek-v4-pro-0813.toml create mode 100644 providers/novita-ai/models/deepseek/deepseek-v4.1-flash.toml create mode 100644 providers/novita-ai/models/deepseek/deepseek_v3.toml delete mode 100644 providers/novita-ai/models/inclusionai/ling-2.6-1t.toml delete mode 100644 providers/novita-ai/models/inclusionai/ling-2.6-flash.toml create mode 100644 providers/novita-ai/models/inclusionai/ling-3.0-flash-fin.toml create mode 100644 providers/novita-ai/models/inclusionai/ling-3.0-flash-sante.toml create mode 100644 providers/novita-ai/models/inclusionai/ling-3.0-flash-vl.toml create mode 100644 providers/novita-ai/models/inclusionai/ling-3.0-flash.toml create mode 100644 providers/novita-ai/models/inclusionai/ling-3.1-flash.toml delete mode 100644 providers/novita-ai/models/inclusionai/ring-2.6-1t.toml delete mode 100644 providers/novita-ai/models/kwaipilot/kat-coder-pro.toml delete mode 100644 providers/novita-ai/models/meta-llama/llama-3-70b-instruct.toml delete mode 100644 providers/novita-ai/models/meta-llama/llama-3-8b-instruct.toml create mode 100644 providers/novita-ai/models/meta-llama/llama-3.2-1b-instruct.toml delete mode 100644 providers/novita-ai/models/meta-llama/llama-3.2-3b-instruct.toml create mode 100644 providers/novita-ai/models/mindai/macaron-v1-tall.toml create mode 100644 providers/novita-ai/models/mindai/macaron-v1-venti.toml delete mode 100644 providers/novita-ai/models/minimax/minimax-m2.1.toml delete mode 100644 providers/novita-ai/models/minimax/minimax-m2.5-highspeed.toml delete mode 100644 providers/novita-ai/models/minimax/minimax-m2.toml create mode 100644 providers/novita-ai/models/minimax/minimax-m3.toml delete mode 100644 providers/novita-ai/models/mistralai/mistral-nemo.toml create mode 100644 providers/novita-ai/models/nvidia/nemotron-3-nano-30b-a3b.toml delete mode 100644 providers/novita-ai/models/openai/gpt-oss-20b.toml delete mode 100644 providers/novita-ai/models/qwen/qwen-2.5-72b-instruct.toml delete mode 100644 providers/novita-ai/models/qwen/qwen2.5-7b-instruct.toml delete mode 100644 providers/novita-ai/models/qwen/qwen2.5-vl-72b-instruct.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-235b-a22b-fp8.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-235b-a22b-instruct-2507.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-235b-a22b-thinking-2507.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-30b-a3b-fp8.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-32b-fp8.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-4b-fp8.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-8b-fp8.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-coder-30b-a3b-instruct.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-coder-480b-a35b-instruct.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-coder-next.toml create mode 100644 providers/novita-ai/models/qwen/qwen3-embedding-8b.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-max.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-next-80b-a3b-instruct.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-next-80b-a3b-thinking.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-instruct.toml create mode 100644 providers/novita-ai/models/qwen/qwen3-reranker-8b.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-vl-235b-a22b-instruct.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-vl-235b-a22b-thinking.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-vl-30b-a3b-thinking.toml delete mode 100644 providers/novita-ai/models/qwen/qwen3-vl-8b-instruct.toml create mode 100644 providers/novita-ai/models/qwen/qwen3.6-27b.toml create mode 100644 providers/novita-ai/models/qwen/qwen3.6-35b-a3b.toml create mode 100644 providers/novita-ai/models/qwen/qwen3.8-2.4t-a95b.toml create mode 100644 providers/novita-ai/models/qwen/qwen3.8-27b.toml create mode 100644 providers/novita-ai/models/qwen/qwen3.8-flash.toml create mode 100644 providers/novita-ai/models/qwen/qwen3.8-max.toml delete mode 100644 providers/novita-ai/models/sao10K/L3-8B-stheno-v3.2.toml delete mode 100644 providers/novita-ai/models/sao10K/l3-70b-euryale-v2.1.toml delete mode 100644 providers/novita-ai/models/sao10K/l3-8b-lunaris.toml delete mode 100644 providers/novita-ai/models/sao10K/l31-70b-euryale-v2.2.toml create mode 100644 providers/novita-ai/models/sao10k/l3-70b-euryale-v2.1.toml create mode 100644 providers/novita-ai/models/stepfun/step-3.7-flash.toml create mode 100644 providers/novita-ai/models/tencent/hy3.toml create mode 100644 providers/novita-ai/models/tencent/hy4-preview.toml delete mode 100644 providers/novita-ai/models/xiaomimimo/mimo-v2-flash.toml delete mode 100644 providers/novita-ai/models/xiaomimimo/mimo-v2-pro.toml create mode 100644 providers/novita-ai/models/xiaomimimo/mimo-v2.5.toml create mode 100644 providers/novita-ai/models/xiaomimimo/mimo-v2.6-flash.toml create mode 100644 providers/novita-ai/models/xiaomimimo/mimo-v2.6-pro.toml delete mode 100644 providers/novita-ai/models/zai-org/glm-4.5.toml create mode 100644 providers/novita-ai/models/zai-org/glm-5.3-flash.toml create mode 100644 providers/novita-ai/models/zai-org/glm-5.3.toml diff --git a/models/alibaba/qwen-mt-plus.toml b/models/alibaba/qwen-mt-plus.toml new file mode 100644 index 00000000000..626d5323e57 --- /dev/null +++ b/models/alibaba/qwen-mt-plus.toml @@ -0,0 +1,22 @@ +# https://www.qwencloud.com/models/qwen-mt-plus +# https://www.alibabacloud.com/help/en/model-studio/qwen-mt-plus +# https://qwenlm.github.io/blog/qwen-mt/ +# https://novita.ai/models/model-detail/qwen-qwen-mt-plus +name = "Qwen MT Plus" +description = "Flagship Qwen translation model with terminology control and support for 92 languages" +family = "qwen" +release_date = "2025-07-24" +last_updated = "2025-07-24" +attachment = false +reasoning = false +tool_call = false +temperature = true +open_weights = false + +[limit] +context = 16_384 +output = 8_192 + +[modalities] +input = ["text"] +output = ["text"] diff --git a/models/alibaba/qwen3-embedding-8b.toml b/models/alibaba/qwen3-embedding-8b.toml new file mode 100644 index 00000000000..63aaa0ec206 --- /dev/null +++ b/models/alibaba/qwen3-embedding-8b.toml @@ -0,0 +1,26 @@ +# https://huggingface.co/Qwen/Qwen3-Embedding-8B +# https://qwenlm.github.io/blog/qwen3-embedding/ +name = "Qwen3 Embedding 8B" +description = "Multilingual text embedding model for retrieval, code search, classification, and clustering" +type = "embedding" +family = "qwen" +release_date = "2025-06-05" +last_updated = "2025-06-05" +attachment = false +reasoning = false +tool_call = false +temperature = false +open_weights = true +license = "Apache-2.0" + +[limit] +context = 32_768 +output = 0 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/Qwen/Qwen3-Embedding-8B" diff --git a/models/alibaba/qwen3-omni-30b-a3b-thinking.toml b/models/alibaba/qwen3-omni-30b-a3b-thinking.toml new file mode 100644 index 00000000000..3300b01ef19 --- /dev/null +++ b/models/alibaba/qwen3-omni-30b-a3b-thinking.toml @@ -0,0 +1,25 @@ +# https://huggingface.co/Qwen/Qwen3-Omni-30B-A3B-Thinking +# https://qwen.ai/blog?id=qwen3-omni +name = "Qwen3 Omni 30B A3B Thinking" +description = "Multimodal reasoning model accepting text, images, video, and audio with text output" +family = "qwen" +release_date = "2025-09-22" +last_updated = "2025-09-22" +attachment = true +reasoning = true +tool_call = true +temperature = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 65_536 +output = 16_384 + +[modalities] +input = ["text", "image", "video", "audio"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/Qwen/Qwen3-Omni-30B-A3B-Thinking" diff --git a/models/alibaba/qwen3-reranker-8b.toml b/models/alibaba/qwen3-reranker-8b.toml new file mode 100644 index 00000000000..4cca889c955 --- /dev/null +++ b/models/alibaba/qwen3-reranker-8b.toml @@ -0,0 +1,26 @@ +# https://huggingface.co/Qwen/Qwen3-Reranker-8B +# https://qwenlm.github.io/blog/qwen3-embedding/ +name = "Qwen3 Reranker 8B" +description = "Multilingual query-document reranking model supporting more than 100 languages" +type = "reranking" +family = "qwen" +release_date = "2025-06-05" +last_updated = "2025-06-05" +attachment = false +reasoning = false +tool_call = false +temperature = false +open_weights = true +license = "Apache-2.0" + +[limit] +context = 32_768 +output = 0 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/Qwen/Qwen3-Reranker-8B" diff --git a/models/alibaba/qwen3-vl-30b-a3b-instruct.toml b/models/alibaba/qwen3-vl-30b-a3b-instruct.toml new file mode 100644 index 00000000000..358a86fbfde --- /dev/null +++ b/models/alibaba/qwen3-vl-30b-a3b-instruct.toml @@ -0,0 +1,25 @@ +# https://huggingface.co/Qwen/Qwen3-VL-30B-A3B-Instruct +# https://github.com/QwenLM/Qwen3-VL (News: 2025.10.4) +name = "Qwen3 VL 30B A3B Instruct" +description = "Open vision-language MoE model for image and video understanding, documents, and visual agents" +family = "qwen" +release_date = "2025-10-04" +last_updated = "2025-10-04" +attachment = true +reasoning = false +tool_call = true +temperature = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 262_144 +output = 32_768 + +[modalities] +input = ["text", "image", "video"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/Qwen/Qwen3-VL-30B-A3B-Instruct" diff --git a/models/apodex/apodex-1.1-mini.toml b/models/apodex/apodex-1.1-mini.toml new file mode 100644 index 00000000000..255c0250cdf --- /dev/null +++ b/models/apodex/apodex-1.1-mini.toml @@ -0,0 +1,25 @@ +# https://www.apodex.com/blog/apodex-1.1-scaling-agentic-intelligence-for-complex-work +# https://huggingface.co/apodex/Apodex-1.1-mini +# Output limit: https://novita.ai/models/model-detail/apodex-apodex-1.1-mini +name = "Apodex 1.1 Mini" +description = "Open-weight 35B MoE model for long-horizon research, forecasting, and tool-assisted work" +release_date = "2026-08-24" +last_updated = "2026-08-24" +attachment = false +reasoning = true +tool_call = true +temperature = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 262_144 +output = 262_144 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/apodex/Apodex-1.1-mini" diff --git a/models/baai/bge-m3.toml b/models/baai/bge-m3.toml new file mode 100644 index 00000000000..0456d46d547 --- /dev/null +++ b/models/baai/bge-m3.toml @@ -0,0 +1,26 @@ +# https://huggingface.co/BAAI/bge-m3 +# https://github.com/FlagOpen/FlagEmbedding/tree/master/research/BGE_M3 +name = "BGE M3" +description = "Multilingual embedding model combining dense, sparse, and multi-vector retrieval" +type = "embedding" +family = "bge" +release_date = "2024-01-30" +last_updated = "2024-01-30" +attachment = false +reasoning = false +tool_call = false +temperature = false +open_weights = true +license = "MIT" + +[limit] +context = 8_192 +output = 0 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/BAAI/bge-m3" diff --git a/models/baai/bge-reranker-v2-m3.toml b/models/baai/bge-reranker-v2-m3.toml new file mode 100644 index 00000000000..cbb218d2287 --- /dev/null +++ b/models/baai/bge-reranker-v2-m3.toml @@ -0,0 +1,25 @@ +# https://huggingface.co/BAAI/bge-reranker-v2-m3 +name = "BGE Reranker V2 M3" +description = "Lightweight multilingual cross-encoder that scores query-document relevance" +type = "reranking" +family = "bge" +release_date = "2024-03-15" +last_updated = "2024-03-15" +attachment = false +reasoning = false +tool_call = false +temperature = false +open_weights = true +license = "Apache-2.0" + +[limit] +context = 8_192 +output = 0 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/BAAI/bge-reranker-v2-m3" diff --git a/models/baichuan/baichuan-m2-32b.toml b/models/baichuan/baichuan-m2-32b.toml new file mode 100644 index 00000000000..302eeb9e2e5 --- /dev/null +++ b/models/baichuan/baichuan-m2-32b.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/baichuan-inc/Baichuan-M2-32B +name = "Baichuan M2 32B" +description = "Open medical model trained for clinical reasoning and patient interaction with on, off, and auto thinking modes" +family = "baichuan" +release_date = "2025-08-13" +last_updated = "2025-08-13" +attachment = false +reasoning = true +tool_call = false +temperature = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 131_072 +output = 131_072 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/baichuan-inc/Baichuan-M2-32B" diff --git a/models/baidu/cobuddy.toml b/models/baidu/cobuddy.toml new file mode 100644 index 00000000000..fe117f40b49 --- /dev/null +++ b/models/baidu/cobuddy.toml @@ -0,0 +1,19 @@ +# https://openrouter.ai/baidu/cobuddy +# https://novita.ai/models/model-detail/baidu-cobuddy +name = "CoBuddy" +description = "Baidu Qianfan model optimized for code generation and tool-using coding agents" +release_date = "2026-05-06" +last_updated = "2026-05-06" +attachment = false +reasoning = true +tool_call = true +temperature = true +open_weights = false + +[limit] +context = 131_072 +output = 65_536 + +[modalities] +input = ["text"] +output = ["text"] diff --git a/models/baidu/ernie-4.5-21b-a3b.toml b/models/baidu/ernie-4.5-21b-a3b.toml new file mode 100644 index 00000000000..8720f5f5488 --- /dev/null +++ b/models/baidu/ernie-4.5-21b-a3b.toml @@ -0,0 +1,25 @@ +# https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-PT +# https://novita.ai/models/model-detail/baidu-ernie-4.5-21B-a3b +name = "ERNIE 4.5 21B A3B" +description = "Open ERNIE text MoE with 21B total and 3B active parameters" +family = "ernie" +release_date = "2025-06-30" +last_updated = "2025-06-30" +attachment = false +reasoning = false +tool_call = true +temperature = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 131_072 +output = 8_000 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-PT" diff --git a/models/deepseek/deepseek-r1-0528-qwen3-8b.toml b/models/deepseek/deepseek-r1-0528-qwen3-8b.toml new file mode 100644 index 00000000000..2301a96718f --- /dev/null +++ b/models/deepseek/deepseek-r1-0528-qwen3-8b.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B +name = "DeepSeek R1 0528 Qwen3 8B" +description = "DeepSeek R1 0528 distilled into Qwen3 8B for mathematics, coding, and reasoning" +family = "deepseek-thinking" +release_date = "2025-05-29" +last_updated = "2025-05-29" +attachment = false +reasoning = true +tool_call = false +temperature = true +open_weights = true +license = "MIT" + +[limit] +context = 131_072 +output = 32_000 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B" diff --git a/models/deepseek/deepseek-r1-0528.toml b/models/deepseek/deepseek-r1-0528.toml new file mode 100644 index 00000000000..3b7de6e614c --- /dev/null +++ b/models/deepseek/deepseek-r1-0528.toml @@ -0,0 +1,25 @@ +# https://huggingface.co/deepseek-ai/DeepSeek-R1-0528 +name = "DeepSeek R1 0528" +description = "May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning" +family = "deepseek-thinking" +release_date = "2025-05-28" +last_updated = "2025-05-28" +attachment = false +reasoning = true +tool_call = true +structured_output = true +temperature = true +open_weights = true +license = "MIT" + +[limit] +context = 163_840 +output = 32_768 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/deepseek-ai/DeepSeek-R1-0528" diff --git a/models/gryphe/mythomax-l2-13b.toml b/models/gryphe/mythomax-l2-13b.toml new file mode 100644 index 00000000000..be17011ef57 --- /dev/null +++ b/models/gryphe/mythomax-l2-13b.toml @@ -0,0 +1,22 @@ +# https://huggingface.co/Gryphe/MythoMax-L2-13b +name = "MythoMax L2 13B" +description = "Llama 2 merge combining MythoLogic and Huginn for creative writing and roleplay" +release_date = "2023-08-10" +last_updated = "2023-08-10" +attachment = false +reasoning = false +tool_call = false +temperature = true +open_weights = true + +[limit] +context = 4_096 +output = 3_200 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/Gryphe/MythoMax-L2-13b" diff --git a/models/inclusionai/ling-3.0-flash-fin.toml b/models/inclusionai/ling-3.0-flash-fin.toml index 921c8c1b14a..06e8ee8aa5d 100644 --- a/models/inclusionai/ling-3.0-flash-fin.toml +++ b/models/inclusionai/ling-3.0-flash-fin.toml @@ -2,15 +2,17 @@ # https://novita.ai/models/model-detail/inclusionai-ling-3.0-flash-fin # https://openrouter.ai/api/v1/models/inclusionai/ling-3.0-flash-fin:free/endpoints # https://vercel.com/changelog/ling-3-0-flash-fin-now-available-on-ai-gateway-for-free +# MIT weights published: https://huggingface.co/inclusionAI/Ling-3.0-flash-Fin (2026-09-03) name = "Ling 3.0 Flash Fin" description = "Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution" family = "ling" release_date = "2026-08-27" -last_updated = "2026-08-27" +last_updated = "2026-09-03" attachment = false reasoning = true tool_call = true -open_weights = false +open_weights = true +license = "MIT" [limit] context = 262_144 @@ -20,6 +22,10 @@ output = 32_768 input = ["text"] output = ["text"] +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/inclusionAI/Ling-3.0-flash-Fin" + [[benchmarks]] name = "FinFIRST" score = 52.85 diff --git a/models/inclusionai/ling-3.0-flash-sante.toml b/models/inclusionai/ling-3.0-flash-sante.toml new file mode 100644 index 00000000000..f8134243c73 --- /dev/null +++ b/models/inclusionai/ling-3.0-flash-sante.toml @@ -0,0 +1,21 @@ +# https://vercel.com/changelog/ling-3-0-flash-sante-is-now-available-on-ai-gateway-for-free +# https://novita.ai/models/model-detail/inclusionai-ling-3.0-flash-sante +# The open Ling backbone does not establish that Sante's medical fine-tune has public weights. +name = "Ling 3.0 Flash Sante" +description = "Health and medicine specialization of Ling 3.0 Flash for medical reasoning and evidence-based retrieval" +family = "ling" +release_date = "2026-09-04" +last_updated = "2026-09-04" +attachment = false +reasoning = true +tool_call = true +temperature = true +open_weights = false + +[limit] +context = 262_144 +output = 32_768 + +[modalities] +input = ["text"] +output = ["text"] diff --git a/models/inclusionai/ling-3.0-flash-vl.toml b/models/inclusionai/ling-3.0-flash-vl.toml new file mode 100644 index 00000000000..22c886acb51 --- /dev/null +++ b/models/inclusionai/ling-3.0-flash-vl.toml @@ -0,0 +1,25 @@ +# https://huggingface.co/inclusionAI/Ling-3.0-flash-VL +# https://openrouter.ai/inclusionai/ling-3.0-flash-vl +name = "Ling 3.0 Flash VL" +description = "Native multimodal hybrid MoE model for image and video understanding, reasoning, and visual agents" +family = "ling" +release_date = "2026-09-10" +last_updated = "2026-09-10" +attachment = true +reasoning = true +tool_call = true +temperature = true +open_weights = true +license = "MIT" + +[limit] +context = 262_144 +output = 32_768 + +[modalities] +input = ["text", "image", "video"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/inclusionAI/Ling-3.0-flash-VL" diff --git a/models/inclusionai/ling-3.0-flash.toml b/models/inclusionai/ling-3.0-flash.toml new file mode 100644 index 00000000000..3a56b857a04 --- /dev/null +++ b/models/inclusionai/ling-3.0-flash.toml @@ -0,0 +1,25 @@ +# https://huggingface.co/inclusionAI/Ling-3.0-flash +# https://openrouter.ai/inclusionai/ling-3.0-flash +name = "Ling 3.0 Flash" +description = "Hybrid-linear MoE reasoning model with 124B total and 5.1B active parameters for efficient agents" +family = "ling" +release_date = "2026-07-23" +last_updated = "2026-07-23" +attachment = false +reasoning = true +tool_call = true +temperature = true +open_weights = true +license = "MIT" + +[limit] +context = 262_144 +output = 32_768 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/inclusionAI/Ling-3.0-flash" diff --git a/models/meta/llama-3.2-1b-instruct.toml b/models/meta/llama-3.2-1b-instruct.toml new file mode 100644 index 00000000000..ff76864155b --- /dev/null +++ b/models/meta/llama-3.2-1b-instruct.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct +name = "Llama 3.2 1B Instruct" +description = "Compact instruction-tuned Llama model for multilingual dialogue and on-device assistants" +family = "llama" +release_date = "2024-09-25" +last_updated = "2024-09-25" +attachment = false +reasoning = false +tool_call = true +temperature = true +open_weights = true +license = "Llama 3.2 Community License" + +[limit] +context = 131_072 +output = 32_000 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct" diff --git a/models/microsoft/wizardlm-2-8x22b.toml b/models/microsoft/wizardlm-2-8x22b.toml new file mode 100644 index 00000000000..5abd23bb865 --- /dev/null +++ b/models/microsoft/wizardlm-2-8x22b.toml @@ -0,0 +1,24 @@ +# https://wizardlm.github.io/WizardLM2/ +# https://huggingface.co/alpindale/WizardLM-2-8x22B (archived original weights) +name = "WizardLM 2 8x22B" +description = "Microsoft WizardLM instruction-tuned Mixtral MoE for multilingual dialogue and complex tasks" +release_date = "2024-04-15" +last_updated = "2024-04-15" +attachment = false +reasoning = false +tool_call = false +temperature = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 65_536 +output = 8_000 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face archive" +url = "https://huggingface.co/alpindale/WizardLM-2-8x22B" diff --git a/models/mindlab/macaron-v1-tall.toml b/models/mindlab/macaron-v1-tall.toml new file mode 100644 index 00000000000..c63fa0457f5 --- /dev/null +++ b/models/mindlab/macaron-v1-tall.toml @@ -0,0 +1,25 @@ +# https://macaron.im/mindlab/research/introducing-macaron-v1 +# https://huggingface.co/mindlab-research/Macaron-V1-Tall +# Output limit: https://novita.ai/models/model-detail/mindai-macaron-v1-tall +name = "Macaron V1 Tall" +description = "Qwen3.6-based Mixture-of-LoRA model with specialists for chat, personal agents, coding, and generative UI" +release_date = "2026-07-21" +last_updated = "2026-07-21" +attachment = false +reasoning = true +tool_call = true +temperature = true +open_weights = true +license = "MIT" + +[limit] +context = 262_144 +output = 32_768 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/mindlab-research/Macaron-V1-Tall" diff --git a/models/mindlab/macaron-v1-venti.toml b/models/mindlab/macaron-v1-venti.toml new file mode 100644 index 00000000000..78f122ead1f --- /dev/null +++ b/models/mindlab/macaron-v1-venti.toml @@ -0,0 +1,25 @@ +# https://macaron.im/mindlab/research/introducing-macaron-v1 +# https://huggingface.co/mindlab-research/Macaron-V1-Venti +# Output limit: https://novita.ai/models/model-detail/mindai-macaron-v1-venti +name = "Macaron V1 Venti" +description = "GLM-5.2-based Mixture-of-LoRA model with specialists for chat, personal agents, coding, and generative UI" +release_date = "2026-07-21" +last_updated = "2026-07-21" +attachment = false +reasoning = true +tool_call = true +temperature = true +open_weights = true +license = "MIT" + +[limit] +context = 1_048_576 +output = 131_072 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/mindlab-research/Macaron-V1-Venti" diff --git a/models/minimax/minimax-m1-80k.toml b/models/minimax/minimax-m1-80k.toml new file mode 100644 index 00000000000..f2f1bd49406 --- /dev/null +++ b/models/minimax/minimax-m1-80k.toml @@ -0,0 +1,25 @@ +# https://huggingface.co/MiniMaxAI/MiniMax-M1-80k +# https://www.minimax.io/news/minimaxm1 +name = "MiniMax M1 80K" +description = "Open hybrid-attention MoE reasoning model with a million-token context and an 80K thinking budget" +family = "minimax" +release_date = "2025-06-17" +last_updated = "2025-06-17" +attachment = false +reasoning = true +tool_call = true +temperature = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 1_000_000 +output = 80_000 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/MiniMaxAI/MiniMax-M1-80k" diff --git a/models/moonshotai/kimi-k2-0905.toml b/models/moonshotai/kimi-k2-0905.toml new file mode 100644 index 00000000000..76cae1e338a --- /dev/null +++ b/models/moonshotai/kimi-k2-0905.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/moonshotai/Kimi-K2-Instruct-0905 +name = "Kimi K2 Instruct 0905" +description = "September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding" +family = "kimi-k2" +release_date = "2025-09-05" +last_updated = "2025-09-05" +attachment = false +reasoning = false +tool_call = true +temperature = true +open_weights = true +license = "Modified MIT" + +[limit] +context = 262_144 +output = 98_304 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/moonshotai/Kimi-K2-Instruct-0905" diff --git a/models/moonshotai/kimi-k2-instruct.toml b/models/moonshotai/kimi-k2-instruct.toml new file mode 100644 index 00000000000..66754546494 --- /dev/null +++ b/models/moonshotai/kimi-k2-instruct.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/moonshotai/Kimi-K2-Instruct +name = "Kimi K2 Instruct" +description = "Non-thinking Kimi K2 MoE instruction model for coding and tool-using agents" +family = "kimi-k2" +release_date = "2025-07-11" +last_updated = "2025-07-11" +attachment = false +reasoning = false +tool_call = true +temperature = true +open_weights = true +license = "Modified MIT" + +[limit] +context = 131_072 +output = 32_768 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/moonshotai/Kimi-K2-Instruct" diff --git a/models/nousresearch/hermes-2-pro-llama-3-8b.toml b/models/nousresearch/hermes-2-pro-llama-3-8b.toml new file mode 100644 index 00000000000..a291d0e9046 --- /dev/null +++ b/models/nousresearch/hermes-2-pro-llama-3-8b.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/NousResearch/Hermes-2-Pro-Llama-3-8B +name = "Hermes 2 Pro Llama 3 8B" +description = "Llama 3 fine-tune optimized for function calling, structured JSON responses, and dialogue" +release_date = "2024-04-30" +last_updated = "2024-04-30" +attachment = false +reasoning = false +tool_call = true +structured_output = true +temperature = true +open_weights = true +license = "Llama 3 Community License" + +[limit] +context = 8_192 +output = 8_192 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/NousResearch/Hermes-2-Pro-Llama-3-8B" diff --git a/models/paddlepaddle/paddleocr-vl.toml b/models/paddlepaddle/paddleocr-vl.toml new file mode 100644 index 00000000000..6190fd2a1a2 --- /dev/null +++ b/models/paddlepaddle/paddleocr-vl.toml @@ -0,0 +1,23 @@ +# https://huggingface.co/PaddlePaddle/PaddleOCR-VL +name = "PaddleOCR VL" +description = "Compact vision-language model for multilingual document parsing, OCR, tables, formulas, and charts" +release_date = "2025-10-16" +last_updated = "2025-10-16" +attachment = true +reasoning = false +tool_call = false +temperature = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 131_072 +output = 16_384 + +[modalities] +input = ["text", "image"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/PaddlePaddle/PaddleOCR-VL" diff --git a/models/sao10k/l3-70b-euryale-v2.1.toml b/models/sao10k/l3-70b-euryale-v2.1.toml new file mode 100644 index 00000000000..aec4c3084fc --- /dev/null +++ b/models/sao10k/l3-70b-euryale-v2.1.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/Sao10K/L3-70B-Euryale-v2.1 +name = "L3 70B Euryale V2.1" +description = "Llama 3 70B fine-tune for creative dialogue and roleplay" +family = "llama" +release_date = "2024-06-11" +last_updated = "2024-06-11" +attachment = false +reasoning = false +tool_call = false +temperature = true +open_weights = true +license = "CC-BY-NC-4.0" + +[limit] +context = 8_192 +output = 8_192 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/Sao10K/L3-70B-Euryale-v2.1" diff --git a/models/sao10k/l3-8b-stheno-v3.2.toml b/models/sao10k/l3-8b-stheno-v3.2.toml new file mode 100644 index 00000000000..3fa3cc5bbf4 --- /dev/null +++ b/models/sao10k/l3-8b-stheno-v3.2.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/Sao10K/L3-8B-Stheno-v3.2 +name = "L3 8B Stheno V3.2" +description = "Llama 3 8B fine-tune for creative writing and roleplay" +family = "llama" +release_date = "2024-06-05" +last_updated = "2024-06-05" +attachment = false +reasoning = false +tool_call = false +temperature = true +open_weights = true +license = "CC-BY-NC-4.0" + +[limit] +context = 8_192 +output = 8_192 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/Sao10K/L3-8B-Stheno-v3.2" diff --git a/models/zhipuai/autoglm-phone-9b-multilingual.toml b/models/zhipuai/autoglm-phone-9b-multilingual.toml new file mode 100644 index 00000000000..8da6a9214e9 --- /dev/null +++ b/models/zhipuai/autoglm-phone-9b-multilingual.toml @@ -0,0 +1,24 @@ +# https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual +name = "AutoGLM Phone 9B Multilingual" +description = "Multilingual visual model for operating smartphone interfaces and mobile agents" +family = "glm" +release_date = "2025-12-10" +last_updated = "2025-12-10" +attachment = true +reasoning = false +tool_call = false +temperature = true +open_weights = true +license = "MIT" + +[limit] +context = 65_536 +output = 65_536 + +[modalities] +input = ["text", "image"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual" diff --git a/providers/novita-ai/models/Sao10K/L3-8B-Stheno-v3.2.toml b/providers/novita-ai/models/Sao10K/L3-8B-Stheno-v3.2.toml new file mode 100644 index 00000000000..ba9d3b4c938 --- /dev/null +++ b/providers/novita-ai/models/Sao10K/L3-8B-Stheno-v3.2.toml @@ -0,0 +1,10 @@ +# https://novita.ai/models/model-detail/Sao10K-L3-8B-Stheno-v3.2 +# Preserve the exact case-sensitive model ID returned by Novita. +base_model = "sao10k/l3-8b-stheno-v3.2" + +[cost] +input = 0.05 +output = 0.05 + +[limit] +output = 32_000 diff --git a/providers/novita-ai/models/apodex/apodex-1.1-mini.toml b/providers/novita-ai/models/apodex/apodex-1.1-mini.toml new file mode 100644 index 00000000000..e046e869dd1 --- /dev/null +++ b/providers/novita-ai/models/apodex/apodex-1.1-mini.toml @@ -0,0 +1,13 @@ +# https://novita.ai/models/model-detail/apodex-apodex-1.1-mini +# Toggle: thinking.type = enabled|disabled +# The public model page and /models both list zero input/output prices on 2026-10-09. +base_model = "apodex/apodex-1.1-mini" +structured_output = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0 +output = 0 diff --git a/providers/novita-ai/models/baai/bge-m3.toml b/providers/novita-ai/models/baai/bge-m3.toml new file mode 100644 index 00000000000..91b3d1736a6 --- /dev/null +++ b/providers/novita-ai/models/baai/bge-m3.toml @@ -0,0 +1,7 @@ +# https://novita.ai/models/model-detail/baai-bge-m3 +# POST /openai/v1/embeddings returns a 1024-dimensional vector; output is not a generative token limit. +base_model = "baai/bge-m3" + +[cost] +input = 0.01 +output = 0.01 diff --git a/providers/novita-ai/models/baai/bge-reranker-v2-m3.toml b/providers/novita-ai/models/baai/bge-reranker-v2-m3.toml new file mode 100644 index 00000000000..253c73be9b8 --- /dev/null +++ b/providers/novita-ai/models/baai/bge-reranker-v2-m3.toml @@ -0,0 +1,10 @@ +# https://novita.ai/models/model-detail/baai-bge-reranker-v2-m3 +# https://novita.ai/docs/api-reference/model-apis-llm-create-rerank +base_model = "baai/bge-reranker-v2-m3" + +[cost] +input = 0.01 +output = 0.01 + +[limit] +context = 8_000 diff --git a/providers/novita-ai/models/baichuan/baichuan-m2-32b.toml b/providers/novita-ai/models/baichuan/baichuan-m2-32b.toml index 61ada86fca5..d81d7c37d03 100644 --- a/providers/novita-ai/models/baichuan/baichuan-m2-32b.toml +++ b/providers/novita-ai/models/baichuan/baichuan-m2-32b.toml @@ -1,24 +1,8 @@ -name = "baichuan-m2-32b" -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -family = "baichuan" -release_date = "2025-08-13" -last_updated = "2025-08-13" -attachment = false +# https://novita.ai/models/model-detail/baichuan-baichuan-m2-32b +# Novita does not advertise reasoning for this route; preserve its non-reasoning serving mode, not the lab's broader thinking capability. +base_model = "baichuan/baichuan-m2-32b" reasoning = false -temperature = true -knowledge = "2024-12" -tool_call = false -structured_output = false -open_weights = true [cost] input = 0.07 output = 0.07 - -[limit] -context = 131072 -output = 131072 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/baidu/cobuddy.toml b/providers/novita-ai/models/baidu/cobuddy.toml new file mode 100644 index 00000000000..f454294346e --- /dev/null +++ b/providers/novita-ai/models/baidu/cobuddy.toml @@ -0,0 +1,12 @@ +# https://novita.ai/models/model-detail/baidu-cobuddy +# Toggle: thinking.type = enabled|disabled (Novita/relay baseline; live checks returned server_overload). +base_model = "baidu/cobuddy" +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.28 +output = 1.13 +cache_read = 0.07 diff --git a/providers/novita-ai/models/baidu/ernie-4.5-21B-a3b-thinking.toml b/providers/novita-ai/models/baidu/ernie-4.5-21B-a3b-thinking.toml deleted file mode 100644 index d8969fa9f63..00000000000 --- a/providers/novita-ai/models/baidu/ernie-4.5-21B-a3b-thinking.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "ERNIE-4.5-21B-A3B-Thinking" -description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use" -family = "ernie" -release_date = "2025-09-19" -last_updated = "2025-09-19" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -knowledge = "2025-03" -tool_call = false -open_weights = true - -[cost] -input = 0.07 -output = 0.28 - -[limit] -context = 131_072 -output = 65_536 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/baidu/ernie-4.5-21B-a3b.toml b/providers/novita-ai/models/baidu/ernie-4.5-21B-a3b.toml index 4329808b555..9914301554c 100644 --- a/providers/novita-ai/models/baidu/ernie-4.5-21B-a3b.toml +++ b/providers/novita-ai/models/baidu/ernie-4.5-21B-a3b.toml @@ -1,14 +1,5 @@ -name = "ERNIE 4.5 21B A3B" -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -family = "ernie" -release_date = "2025-06-30" -last_updated = "2025-06-30" -attachment = false -reasoning = false -temperature = true -knowledge = "2025-03" -tool_call = true -open_weights = true +# https://novita.ai/models/model-detail/baidu-ernie-4.5-21B-a3b +base_model = "baidu/ernie-4.5-21b-a3b" [cost] input = 0.07 @@ -16,8 +7,3 @@ output = 0.28 [limit] context = 120_000 -output = 8_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/baidu/ernie-4.5-300b-a47b-paddle.toml b/providers/novita-ai/models/baidu/ernie-4.5-300b-a47b-paddle.toml deleted file mode 100644 index b53b21ad54a..00000000000 --- a/providers/novita-ai/models/baidu/ernie-4.5-300b-a47b-paddle.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "ERNIE 4.5 300B A47B" -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -release_date = "2025-06-30" -last_updated = "2025-06-30" -attachment = false -reasoning = false -temperature = true -tool_call = false -structured_output = true -open_weights = true - -[cost] -input = 0.28 -output = 1.1 - -[limit] -context = 123_000 -output = 12_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/baidu/ernie-4.5-vl-28b-a3b-thinking.toml b/providers/novita-ai/models/baidu/ernie-4.5-vl-28b-a3b-thinking.toml deleted file mode 100644 index 15b24e7dc0e..00000000000 --- a/providers/novita-ai/models/baidu/ernie-4.5-vl-28b-a3b-thinking.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "ERNIE-4.5-VL-28B-A3B-Thinking" -description = "Multimodal reasoning model for visual analysis, planning, and tool use" -release_date = "2025-11-26" -last_updated = "2025-11-26" -attachment = true -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.39 -output = 0.39 - -[limit] -context = 131_072 -output = 65_536 - -[modalities] -input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/baidu/ernie-4.5-vl-28b-a3b.toml b/providers/novita-ai/models/baidu/ernie-4.5-vl-28b-a3b.toml deleted file mode 100644 index 6a9968ee3f2..00000000000 --- a/providers/novita-ai/models/baidu/ernie-4.5-vl-28b-a3b.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "ERNIE 4.5 VL 28B A3B" -description = "Multimodal reasoning model for visual analysis, planning, and tool use" -release_date = "2025-06-30" -last_updated = "2026-06-14" -attachment = true -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true -open_weights = true - -[cost] -input = 0.14 -output = 0.56 - -[limit] -context = 30_000 -output = 8_000 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/baidu/ernie-4.5-vl-424b-a47b.toml b/providers/novita-ai/models/baidu/ernie-4.5-vl-424b-a47b.toml deleted file mode 100644 index 2f7125b0bcc..00000000000 --- a/providers/novita-ai/models/baidu/ernie-4.5-vl-424b-a47b.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "ERNIE 4.5 VL 424B A47B" -description = "Multimodal reasoning model for visual analysis, planning, and tool use" -release_date = "2025-06-30" -last_updated = "2025-06-30" -attachment = true -reasoning = true -reasoning_options = [] -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.42 -output = 1.25 - -[limit] -context = 123_000 -output = 16_000 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-ocr-2.toml b/providers/novita-ai/models/deepseek/deepseek-ocr-2.toml index c687f5c2da1..881c19a23c3 100644 --- a/providers/novita-ai/models/deepseek/deepseek-ocr-2.toml +++ b/providers/novita-ai/models/deepseek/deepseek-ocr-2.toml @@ -1,20 +1,6 @@ -name = "deepseek/deepseek-ocr-2" -description = "OCR model for extracting structured text from documents and screenshots" -release_date = "2026-01-27" -last_updated = "2026-01-27" -attachment = true -reasoning = false -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/deepseek-deepseek-ocr-2 +base_model = "deepseek/deepseek-ocr-2" [cost] input = 0.03 output = 0.03 - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-ocr.toml b/providers/novita-ai/models/deepseek/deepseek-ocr.toml deleted file mode 100644 index 56f6fc709bf..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-ocr.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "DeepSeek-OCR" -description = "OCR model for extracting structured text from documents and screenshots" -release_date = "2025-10-24" -last_updated = "2025-10-24" -attachment = true -reasoning = false -temperature = true -tool_call = false -structured_output = true -open_weights = true - -[cost] -input = 0.03 -output = 0.03 - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-prover-v2-671b.toml b/providers/novita-ai/models/deepseek/deepseek-prover-v2-671b.toml deleted file mode 100644 index 06aac3d3e1f..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-prover-v2-671b.toml +++ /dev/null @@ -1,21 +0,0 @@ -name = "Deepseek Prover V2 671B" -description = "Flagship DeepSeek model for coding, reasoning, and agentic work" -release_date = "2025-04-30" -last_updated = "2025-04-30" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.7 -output = 2.5 - -[limit] -context = 160_000 -output = 160_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-r1-0528-qwen3-8b.toml b/providers/novita-ai/models/deepseek/deepseek-r1-0528-qwen3-8b.toml index a3c7e763158..b4612543ae8 100644 --- a/providers/novita-ai/models/deepseek/deepseek-r1-0528-qwen3-8b.toml +++ b/providers/novita-ai/models/deepseek/deepseek-r1-0528-qwen3-8b.toml @@ -1,13 +1,7 @@ -name = "DeepSeek R1 0528 Qwen3 8B" -description = "DeepSeek reasoning model for multi-step analysis, math, coding, and tools" -release_date = "2025-05-29" -last_updated = "2025-05-29" -attachment = false -reasoning = true +# https://novita.ai/models/model-detail/deepseek-deepseek-r1-0528-qwen3-8b +# Always-on distilled reasoning model; live availability checks returned 503, not MODEL_NOT_FOUND. +base_model = "deepseek/deepseek-r1-0528-qwen3-8b" reasoning_options = [] -temperature = true -tool_call = false -open_weights = true [cost] input = 0.06 @@ -15,8 +9,3 @@ output = 0.09 [limit] context = 128_000 -output = 32_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-r1-0528.toml b/providers/novita-ai/models/deepseek/deepseek-r1-0528.toml index b84abfabbca..776c6a472ac 100644 --- a/providers/novita-ai/models/deepseek/deepseek-r1-0528.toml +++ b/providers/novita-ai/models/deepseek/deepseek-r1-0528.toml @@ -1,25 +1,9 @@ -name = "DeepSeek R1 0528" -description = "DeepSeek reasoning model for multi-step analysis, math, coding, and tools" -family = "deepseek-thinking" -release_date = "2025-05-28" -last_updated = "2025-05-28" -attachment = false -reasoning = true +# https://novita.ai/models/model-detail/deepseek-deepseek-r1-0528 +# Always-on reasoning is emitted as in ordinary content on this route; no separate reasoning side channel was observed. +base_model = "deepseek/deepseek-r1-0528" reasoning_options = [] -temperature = true -knowledge = "2024-07" -tool_call = true -structured_output = true -open_weights = true [cost] input = 0.7 output = 2.5 cache_read = 0.35 -[limit] -context = 163_840 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-r1-distill-llama-70b.toml b/providers/novita-ai/models/deepseek/deepseek-r1-distill-llama-70b.toml deleted file mode 100644 index 500a14e21e9..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-r1-distill-llama-70b.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "DeepSeek R1 Distill LLama 70B" -description = "DeepSeek reasoning model for multi-step analysis, math, coding, and tools" -family = "deepseek-thinking" -release_date = "2025-01-27" -last_updated = "2025-01-27" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = false -structured_output = true -open_weights = true - -[cost] -input = 0.8 -output = 0.8 - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-r1-distill-qwen-14b.toml b/providers/novita-ai/models/deepseek/deepseek-r1-distill-qwen-14b.toml deleted file mode 100644 index a05d9cbaa66..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-r1-distill-qwen-14b.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "DeepSeek R1 Distill Qwen 14B" -description = "Qwen reasoning model for deliberate problem solving, math, and coding" -family = "deepseek-thinking" -release_date = "2025-01-20" -last_updated = "2025-01-20" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.15 -output = 0.15 - -[limit] -context = 32_768 -output = 16_384 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-r1-distill-qwen-32b.toml b/providers/novita-ai/models/deepseek/deepseek-r1-distill-qwen-32b.toml deleted file mode 100644 index 3b78039a9dd..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-r1-distill-qwen-32b.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "DeepSeek R1 Distill Qwen 32B" -description = "Qwen reasoning model for deliberate problem solving, math, and coding" -family = "deepseek-thinking" -release_date = "2025-01-20" -last_updated = "2025-01-20" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.3 -output = 0.3 - -[limit] -context = 64_000 -output = 32_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-r1-turbo.toml b/providers/novita-ai/models/deepseek/deepseek-r1-turbo.toml index 23da112f12b..6ca16fdf0ad 100644 --- a/providers/novita-ai/models/deepseek/deepseek-r1-turbo.toml +++ b/providers/novita-ai/models/deepseek/deepseek-r1-turbo.toml @@ -1,13 +1,9 @@ -name = "DeepSeek R1 (Turbo) " -description = "DeepSeek reasoning model for multi-step analysis, math, coding, and tools" -release_date = "2025-03-05" -last_updated = "2025-03-05" -attachment = false -reasoning = true +# https://novita.ai/models/model-detail/deepseek-deepseek-r1-turbo +# thinking.type and separate_reasoning did not change the inline reasoning observed on this route. +base_model = "deepseek/deepseek-r1" +name = "DeepSeek R1 (Turbo)" +structured_output = true reasoning_options = [] -temperature = true -tool_call = true -open_weights = true [cost] input = 0.7 @@ -16,7 +12,3 @@ output = 2.5 [limit] context = 64_000 output = 16_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-v3-0324.toml b/providers/novita-ai/models/deepseek/deepseek-v3-0324.toml deleted file mode 100644 index 4a4c1961851..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-v3-0324.toml +++ /dev/null @@ -1,25 +0,0 @@ -name = "DeepSeek V3 0324" -description = "DeepSeek chat model for instruction following, coding, and analysis" -family = "deepseek" -release_date = "2025-03-25" -last_updated = "2025-03-25" -attachment = false -reasoning = false -temperature = true -knowledge = "2024-07" -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.27 -output = 1.12 -cache_read = 0.135 - -[limit] -context = 163_840 -output = 163_840 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-v3-turbo.toml b/providers/novita-ai/models/deepseek/deepseek-v3-turbo.toml deleted file mode 100644 index 951e07096ee..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-v3-turbo.toml +++ /dev/null @@ -1,21 +0,0 @@ -name = "DeepSeek V3 (Turbo) " -description = "Fast DeepSeek model for efficient chat, coding help, and agent loops" -release_date = "2025-03-05" -last_updated = "2025-03-05" -attachment = false -reasoning = false -temperature = true -tool_call = true -open_weights = true - -[cost] -input = 0.4 -output = 1.3 - -[limit] -context = 64_000 -output = 16_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-v3.1-terminus.toml b/providers/novita-ai/models/deepseek/deepseek-v3.1-terminus.toml deleted file mode 100644 index f12185cccab..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-v3.1-terminus.toml +++ /dev/null @@ -1,27 +0,0 @@ -name = "Deepseek V3.1 Terminus" -description = "DeepSeek chat model for instruction following, coding, and analysis" -family = "deepseek" -release_date = "2025-09-22" -last_updated = "2025-09-22" -attachment = false -reasoning = true -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[[reasoning_options]] -type = "toggle" - -[cost] -input = 0.27 -output = 1.0 -cache_read = 0.135 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-v3.1.toml b/providers/novita-ai/models/deepseek/deepseek-v3.1.toml deleted file mode 100644 index 68208d89a44..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-v3.1.toml +++ /dev/null @@ -1,27 +0,0 @@ -name = "DeepSeek V3.1" -description = "DeepSeek chat model for instruction following, coding, and analysis" -family = "deepseek" -release_date = "2025-08-21" -last_updated = "2025-08-21" -attachment = false -reasoning = true -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[[reasoning_options]] -type = "toggle" - -[cost] -input = 0.27 -output = 1.0 -cache_read = 0.135 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-v3.2-exp.toml b/providers/novita-ai/models/deepseek/deepseek-v3.2-exp.toml deleted file mode 100644 index 7f4e845ee86..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-v3.2-exp.toml +++ /dev/null @@ -1,25 +0,0 @@ -name = "Deepseek V3.2 Exp" -description = "DeepSeek chat model for instruction following, coding, and analysis" -release_date = "2025-09-29" -last_updated = "2025-09-29" -attachment = false -reasoning = true -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[[reasoning_options]] -type = "toggle" - -[cost] -input = 0.27 -output = 0.41 - -[limit] -context = 163_840 -output = 65_536 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-v3.2.toml b/providers/novita-ai/models/deepseek/deepseek-v3.2.toml deleted file mode 100644 index 88862eac490..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-v3.2.toml +++ /dev/null @@ -1,28 +0,0 @@ -name = "Deepseek V3.2" -description = "DeepSeek chat model for instruction following, coding, and analysis" -family = "deepseek" -release_date = "2025-12-01" -last_updated = "2025-12-01" -attachment = false -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.269 -output = 0.4 -cache_read = 0.1345 - -[limit] -context = 163_840 -output = 65_536 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-flash-0731.toml b/providers/novita-ai/models/deepseek/deepseek-v4-flash-0731.toml new file mode 100644 index 00000000000..6bcc65c1fb8 --- /dev/null +++ b/providers/novita-ai/models/deepseek/deepseek-v4-flash-0731.toml @@ -0,0 +1,17 @@ +# https://novita.ai/models/model-detail/deepseek-deepseek-v4-flash-0731 +# Toggle: thinking.type = enabled|disabled +# Effort: reasoning_effort = high|max (native baseline; low accepted as an alias). +base_model = "deepseek/deepseek-v4-flash-0731" +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["high", "max"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.44 +output = 1.32 +cache_read = 0.028 + +[limit] +context = 1_048_576 +output = 393_216 diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml b/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml index 09253a5a834..687891efaf9 100644 --- a/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml +++ b/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml @@ -1,5 +1,9 @@ +# https://novita.ai/models/model-detail/deepseek-deepseek-v4-flash +# Toggle: thinking.type = enabled|disabled +# Effort: reasoning_effort = high|max (native baseline; low accepted as an alias). +# Explicit effort=high currently returns MODEL_NOT_FOUND, while default, low, and max requests work; retain the route. base_model = "deepseek/deepseek-v4-flash" -reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["minimal", "low", "medium", "high", "xhigh"] }] +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["high", "max"] }] [interleaved] field = "reasoning_content" diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-pro-0813.toml b/providers/novita-ai/models/deepseek/deepseek-v4-pro-0813.toml new file mode 100644 index 00000000000..b0f7129b90a --- /dev/null +++ b/providers/novita-ai/models/deepseek/deepseek-v4-pro-0813.toml @@ -0,0 +1,17 @@ +# https://novita.ai/models/model-detail/deepseek-deepseek-v4-pro-0813 +# Toggle: thinking.type = enabled|disabled +# Effort: reasoning_effort = high|max (native baseline; low accepted as an alias). +base_model = "deepseek/deepseek-v4-pro-0813" +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["high", "max"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 1.32 +output = 3.96 +cache_read = 0.044 + +[limit] +context = 1_048_576 +output = 393_216 diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-pro.toml b/providers/novita-ai/models/deepseek/deepseek-v4-pro.toml index eb84d4bdfd5..1b0c165e35e 100644 --- a/providers/novita-ai/models/deepseek/deepseek-v4-pro.toml +++ b/providers/novita-ai/models/deepseek/deepseek-v4-pro.toml @@ -1,5 +1,8 @@ +# https://novita.ai/models/model-detail/deepseek-deepseek-v4-pro +# Toggle: thinking.type = enabled|disabled +# Effort: reasoning_effort = high|max (native baseline; low accepted as an alias). base_model = "deepseek/deepseek-v4-pro" -reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "medium", "high", "xhigh"] }] +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["high", "max"] }] [interleaved] field = "reasoning_content" diff --git a/providers/novita-ai/models/deepseek/deepseek-v4.1-flash.toml b/providers/novita-ai/models/deepseek/deepseek-v4.1-flash.toml new file mode 100644 index 00000000000..bde50d02b97 --- /dev/null +++ b/providers/novita-ai/models/deepseek/deepseek-v4.1-flash.toml @@ -0,0 +1,17 @@ +# https://novita.ai/models/model-detail/deepseek-deepseek-v4.1-flash +# Toggle: thinking.type = enabled|disabled +# Effort: reasoning_effort = low|high|max; invalid values are rejected. +base_model = "deepseek/deepseek-v4.1-flash" +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "high", "max"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.3 +output = 1.2 +cache_read = 0.006 + +[limit] +context = 1_048_576 +output = 393_216 diff --git a/providers/novita-ai/models/deepseek/deepseek_v3.toml b/providers/novita-ai/models/deepseek/deepseek_v3.toml new file mode 100644 index 00000000000..85c991ac7ce --- /dev/null +++ b/providers/novita-ai/models/deepseek/deepseek_v3.toml @@ -0,0 +1,11 @@ +# https://novita.ai/docs/api-reference/model-apis-llm-list-models +# Exact legacy ID returned by /models and verified with a successful chat request on 2026-10-09. +base_model = "deepseek/deepseek-v3" + +[cost] +input = 0.89 +output = 0.89 + +[limit] +context = 64_000 +output = 16_000 diff --git a/providers/novita-ai/models/google/gemma-3-12b-it.toml b/providers/novita-ai/models/google/gemma-3-12b-it.toml index 01e59f8e1ed..afa1d7bcf2f 100644 --- a/providers/novita-ai/models/google/gemma-3-12b-it.toml +++ b/providers/novita-ai/models/google/gemma-3-12b-it.toml @@ -1,22 +1,9 @@ -name = "Gemma 3 12B" -description = "Open Gemma instruction model for efficient chat and self-hosted deployments" -family = "gemma" -release_date = "2025-03-13" -last_updated = "2025-03-13" -attachment = true -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/google-gemma-3-12b-it +base_model = "google/gemma-3-12b-it" [cost] input = 0.05 output = 0.1 [limit] -context = 131_072 output = 8_192 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/google/gemma-3-27b-it.toml b/providers/novita-ai/models/google/gemma-3-27b-it.toml index c04dda55982..752e30705cc 100644 --- a/providers/novita-ai/models/google/gemma-3-27b-it.toml +++ b/providers/novita-ai/models/google/gemma-3-27b-it.toml @@ -1,13 +1,5 @@ -name = "Gemma 3 27B" -description = "Open Gemma instruction model for efficient chat and self-hosted deployments" -family = "gemma" -release_date = "2025-03-25" -last_updated = "2025-03-25" -attachment = true -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/google-gemma-3-27b-it +base_model = "google/gemma-3-27b-it" [cost] input = 0.119 @@ -16,7 +8,3 @@ output = 0.2 [limit] context = 98_304 output = 16_384 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/google/gemma-4-26b-a4b-it.toml b/providers/novita-ai/models/google/gemma-4-26b-a4b-it.toml index 1f22792e823..99bf4aee803 100644 --- a/providers/novita-ai/models/google/gemma-4-26b-a4b-it.toml +++ b/providers/novita-ai/models/google/gemma-4-26b-a4b-it.toml @@ -1,24 +1,14 @@ -name = "Gemma 4 26B A4B" -description = "Open Gemma instruction model for efficient chat and self-hosted deployments" -family = "gemma" -release_date = "2026-04-02" -last_updated = "2026-04-02" -attachment = true -reasoning = true +# https://novita.ai/models/model-detail/google-gemma-4-26b-a4b-it +# Toggle: thinking.type = enabled|disabled +base_model = "google/gemma-4-26b-a4b-it" reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -open_weights = true + +[interleaved] +field = "reasoning_content" [cost] input = 0.13 output = 0.4 [limit] -context = 262_144 output = 131_072 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/google/gemma-4-31b-it.toml b/providers/novita-ai/models/google/gemma-4-31b-it.toml index 6543690e3a0..e27eadfee64 100644 --- a/providers/novita-ai/models/google/gemma-4-31b-it.toml +++ b/providers/novita-ai/models/google/gemma-4-31b-it.toml @@ -1,24 +1,14 @@ -name = "Gemma 4 31B" -description = "Open Gemma instruction model for efficient chat and self-hosted deployments" -family = "gemma" -release_date = "2026-04-02" -last_updated = "2026-04-02" -attachment = true -reasoning = true +# https://novita.ai/models/model-detail/google-gemma-4-31b-it +# Toggle: thinking.type = enabled|disabled +base_model = "google/gemma-4-31b-it" reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -open_weights = true + +[interleaved] +field = "reasoning_content" [cost] input = 0.14 output = 0.4 [limit] -context = 262_144 output = 131_072 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/gryphe/mythomax-l2-13b.toml b/providers/novita-ai/models/gryphe/mythomax-l2-13b.toml index 718c181908f..f2a5ac89c4a 100644 --- a/providers/novita-ai/models/gryphe/mythomax-l2-13b.toml +++ b/providers/novita-ai/models/gryphe/mythomax-l2-13b.toml @@ -1,21 +1,6 @@ -name = "Mythomax L2 13B" -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -release_date = "2024-04-25" -last_updated = "2024-04-25" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/gryphe-mythomax-l2-13b +base_model = "gryphe/mythomax-l2-13b" [cost] input = 0.09 output = 0.09 - -[limit] -context = 4_096 -output = 3_200 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/inclusionai/ling-2.6-1t.toml b/providers/novita-ai/models/inclusionai/ling-2.6-1t.toml deleted file mode 100644 index f41fe3ab980..00000000000 --- a/providers/novita-ai/models/inclusionai/ling-2.6-1t.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "Ling-2.6-1T" -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -family = "ling" -release_date = "2026-04-23" -last_updated = "2026-06-29" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.3 -output = 2.5 -cache_read = 0.06 - -[limit] -context = 262_144 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/inclusionai/ling-2.6-flash.toml b/providers/novita-ai/models/inclusionai/ling-2.6-flash.toml deleted file mode 100644 index ae3dd2c3c31..00000000000 --- a/providers/novita-ai/models/inclusionai/ling-2.6-flash.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "Ling-2.6-flash" -description = "Efficient model for low-latency assistance, extraction, and routine automation" -family = "ling" -release_date = "2026-04-24" -last_updated = "2026-04-24" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.1 -output = 0.3 -cache_read = 0.02 - -[limit] -context = 262_144 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/inclusionai/ling-3.0-flash-fin.toml b/providers/novita-ai/models/inclusionai/ling-3.0-flash-fin.toml new file mode 100644 index 00000000000..25b188b41c3 --- /dev/null +++ b/providers/novita-ai/models/inclusionai/ling-3.0-flash-fin.toml @@ -0,0 +1,12 @@ +# https://novita.ai/models/model-detail/inclusionai-ling-3.0-flash-fin +# Toggle: thinking.type = enabled|disabled +base_model = "inclusionai/ling-3.0-flash-fin" +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.075 +output = 0.22 +cache_read = 0.015 diff --git a/providers/novita-ai/models/inclusionai/ling-3.0-flash-sante.toml b/providers/novita-ai/models/inclusionai/ling-3.0-flash-sante.toml new file mode 100644 index 00000000000..d542eeea829 --- /dev/null +++ b/providers/novita-ai/models/inclusionai/ling-3.0-flash-sante.toml @@ -0,0 +1,12 @@ +# https://novita.ai/models/model-detail/inclusionai-ling-3.0-flash-sante +# Toggle: thinking.type = enabled|disabled +base_model = "inclusionai/ling-3.0-flash-sante" +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.075 +output = 0.22 +cache_read = 0.015 diff --git a/providers/novita-ai/models/inclusionai/ling-3.0-flash-vl.toml b/providers/novita-ai/models/inclusionai/ling-3.0-flash-vl.toml new file mode 100644 index 00000000000..a5363497486 --- /dev/null +++ b/providers/novita-ai/models/inclusionai/ling-3.0-flash-vl.toml @@ -0,0 +1,12 @@ +# https://novita.ai/models/model-detail/inclusionai-ling-3.0-flash-vl +# Toggle: thinking.type = enabled|disabled +base_model = "inclusionai/ling-3.0-flash-vl" +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.075 +output = 0.22 +cache_read = 0.015 diff --git a/providers/novita-ai/models/inclusionai/ling-3.0-flash.toml b/providers/novita-ai/models/inclusionai/ling-3.0-flash.toml new file mode 100644 index 00000000000..82560679a66 --- /dev/null +++ b/providers/novita-ai/models/inclusionai/ling-3.0-flash.toml @@ -0,0 +1,12 @@ +# https://novita.ai/models/model-detail/inclusionai-ling-3.0-flash +# Toggle: thinking.type = enabled|disabled +base_model = "inclusionai/ling-3.0-flash" +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.06 +output = 0.18 +cache_read = 0.012 diff --git a/providers/novita-ai/models/inclusionai/ling-3.1-flash.toml b/providers/novita-ai/models/inclusionai/ling-3.1-flash.toml new file mode 100644 index 00000000000..56e166f3364 --- /dev/null +++ b/providers/novita-ai/models/inclusionai/ling-3.1-flash.toml @@ -0,0 +1,12 @@ +# https://novita.ai/models/model-detail/inclusionai-ling-3.1-flash +# Toggle: thinking.type = enabled|disabled (Ling peer baseline; live checks returned server_overload). +# The public model page and /models both list zero input/output prices on 2026-10-09. +base_model = "inclusionai/ling-3.1-flash" +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0 +output = 0 diff --git a/providers/novita-ai/models/inclusionai/ring-2.6-1t.toml b/providers/novita-ai/models/inclusionai/ring-2.6-1t.toml deleted file mode 100644 index 4555e9bb9cd..00000000000 --- a/providers/novita-ai/models/inclusionai/ring-2.6-1t.toml +++ /dev/null @@ -1,25 +0,0 @@ -name = "Ring-2.6-1T" -description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use" -family = "ring" -release_date = "2026-05-08" -last_updated = "2026-05-27" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true -structured_output = true -open_weights = false - -[cost] -input = 0.3 -output = 2.5 -cache_read = 0.06 - -[limit] -context = 262_144 -output = 65_536 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/kwaipilot/kat-coder-pro.toml b/providers/novita-ai/models/kwaipilot/kat-coder-pro.toml deleted file mode 100644 index 5cab1ba8796..00000000000 --- a/providers/novita-ai/models/kwaipilot/kat-coder-pro.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "Kat Coder Pro" -description = "Coding model for repository understanding, refactors, and agentic engineering tasks" -release_date = "2026-01-05" -last_updated = "2026-01-05" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.3 -output = 1.2 -cache_read = 0.06 - -[limit] -context = 256_000 -output = 128_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/meta-llama/llama-3-70b-instruct.toml b/providers/novita-ai/models/meta-llama/llama-3-70b-instruct.toml deleted file mode 100644 index 5f90f05b22d..00000000000 --- a/providers/novita-ai/models/meta-llama/llama-3-70b-instruct.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "Llama3 70B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-04-25" -last_updated = "2024-04-25" -attachment = false -reasoning = false -temperature = true -tool_call = false -structured_output = true -open_weights = true - -[cost] -input = 0.51 -output = 0.74 - -[limit] -context = 8_192 -output = 8_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/meta-llama/llama-3-8b-instruct.toml b/providers/novita-ai/models/meta-llama/llama-3-8b-instruct.toml deleted file mode 100644 index 8c7143f0b68..00000000000 --- a/providers/novita-ai/models/meta-llama/llama-3-8b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Llama 3 8B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-04-25" -last_updated = "2024-04-25" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.04 -output = 0.04 - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/meta-llama/llama-3.1-8b-instruct.toml b/providers/novita-ai/models/meta-llama/llama-3.1-8b-instruct.toml index c04dc71995b..bacbb7670ad 100644 --- a/providers/novita-ai/models/meta-llama/llama-3.1-8b-instruct.toml +++ b/providers/novita-ai/models/meta-llama/llama-3.1-8b-instruct.toml @@ -1,13 +1,6 @@ -name = "Llama 3.1 8B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-07-24" -last_updated = "2024-07-24" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/meta-llama-llama-3.1-8b-instruct +base_model = "meta/llama-3.1-8b-instruct" +structured_output = true [cost] input = 0.02 @@ -16,7 +9,3 @@ output = 0.05 [limit] context = 16_384 output = 16_384 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/meta-llama/llama-3.2-1b-instruct.toml b/providers/novita-ai/models/meta-llama/llama-3.2-1b-instruct.toml new file mode 100644 index 00000000000..8a5e52c703a --- /dev/null +++ b/providers/novita-ai/models/meta-llama/llama-3.2-1b-instruct.toml @@ -0,0 +1,11 @@ +# https://novita.ai/docs/api-reference/model-apis-llm-list-models +# Exact ID, pricing, and limits returned by /models on 2026-10-09; live inference returned 503, not MODEL_NOT_FOUND. +base_model = "meta/llama-3.2-1b-instruct" +structured_output = true + +[cost] +input = 0.02 +output = 0.02 + +[limit] +context = 131_000 diff --git a/providers/novita-ai/models/meta-llama/llama-3.2-3b-instruct.toml b/providers/novita-ai/models/meta-llama/llama-3.2-3b-instruct.toml deleted file mode 100644 index ec9e35234e4..00000000000 --- a/providers/novita-ai/models/meta-llama/llama-3.2-3b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Llama 3.2 3B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-09-18" -last_updated = "2024-09-18" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.03 -output = 0.05 - -[limit] -context = 32_768 -output = 32_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/meta-llama/llama-3.3-70b-instruct.toml b/providers/novita-ai/models/meta-llama/llama-3.3-70b-instruct.toml index c1e6260cb4b..2fc99415476 100644 --- a/providers/novita-ai/models/meta-llama/llama-3.3-70b-instruct.toml +++ b/providers/novita-ai/models/meta-llama/llama-3.3-70b-instruct.toml @@ -1,23 +1,11 @@ -name = "Llama 3.3 70B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-12-07" -last_updated = "2024-12-07" -attachment = false -reasoning = false -temperature = true -knowledge = "2023-12" -tool_call = true -open_weights = true +# https://novita.ai/models/model-detail/meta-llama-llama-3.3-70b-instruct +base_model = "meta/llama-3.3-70b-instruct" +structured_output = true [cost] input = 0.135 output = 0.4 [limit] -context = 131_072 -output = 120_000 - -[modalities] -input = ["text"] -output = ["text"] +context = 12_288 +output = 12_288 diff --git a/providers/novita-ai/models/meta-llama/llama-4-maverick-17b-128e-instruct-fp8.toml b/providers/novita-ai/models/meta-llama/llama-4-maverick-17b-128e-instruct-fp8.toml index c4aa16dcef9..51c10b0c1f9 100644 --- a/providers/novita-ai/models/meta-llama/llama-4-maverick-17b-128e-instruct-fp8.toml +++ b/providers/novita-ai/models/meta-llama/llama-4-maverick-17b-128e-instruct-fp8.toml @@ -1,12 +1,6 @@ -name = "Llama 4 Maverick Instruct" -description = "Open multimodal Llama model for strong reasoning and fast responses" -release_date = "2025-04-06" -last_updated = "2025-04-06" -attachment = true -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/meta-llama-llama-4-maverick-17b-128e-instruct-fp8 +base_model = "meta/llama-4-maverick-17b-instruct" +structured_output = true [cost] input = 0.27 @@ -15,7 +9,3 @@ output = 0.85 [limit] context = 1_048_576 output = 8_192 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/meta-llama/llama-4-scout-17b-16e-instruct.toml b/providers/novita-ai/models/meta-llama/llama-4-scout-17b-16e-instruct.toml index 29aa8fa750d..261db8db13f 100644 --- a/providers/novita-ai/models/meta-llama/llama-4-scout-17b-16e-instruct.toml +++ b/providers/novita-ai/models/meta-llama/llama-4-scout-17b-16e-instruct.toml @@ -1,12 +1,5 @@ -name = "Llama 4 Scout Instruct" -description = "Open multimodal Llama model for long-context analysis and efficient agents" -release_date = "2025-04-06" -last_updated = "2025-04-06" -attachment = true -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/meta-llama-llama-4-scout-17b-16e-instruct +base_model = "meta/llama-4-scout-17b-instruct" [cost] input = 0.18 @@ -15,7 +8,3 @@ output = 0.59 [limit] context = 131_072 output = 131_072 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/microsoft/wizardlm-2-8x22b.toml b/providers/novita-ai/models/microsoft/wizardlm-2-8x22b.toml index 37207587026..cfd990f83c5 100644 --- a/providers/novita-ai/models/microsoft/wizardlm-2-8x22b.toml +++ b/providers/novita-ai/models/microsoft/wizardlm-2-8x22b.toml @@ -1,12 +1,6 @@ -name = "Wizardlm 2 8x22B" -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -release_date = "2024-04-24" -last_updated = "2024-04-24" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/microsoft-wizardlm-2-8x22b +base_model = "microsoft/wizardlm-2-8x22b" +structured_output = true [cost] input = 0.62 @@ -14,8 +8,3 @@ output = 0.62 [limit] context = 65_535 -output = 8_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/mindai/macaron-v1-tall.toml b/providers/novita-ai/models/mindai/macaron-v1-tall.toml new file mode 100644 index 00000000000..390c9f5da46 --- /dev/null +++ b/providers/novita-ai/models/mindai/macaron-v1-tall.toml @@ -0,0 +1,14 @@ +# https://novita.ai/models/model-detail/mindai-macaron-v1-tall +# Toggle: reasoning_effort = none|high (off|on); this field, not thinking.type, suppresses reasoning. +# Other accepted effort values were not demonstrated to be distinct depth controls, so only the verified binary control is represented. +# thinking.type=disabled does not suppress reasoning on this route; thinking_budget is not honored. +base_model = "mindlab/macaron-v1-tall" +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.2475 +output = 1.43 +cache_read = 0.044 diff --git a/providers/novita-ai/models/mindai/macaron-v1-venti.toml b/providers/novita-ai/models/mindai/macaron-v1-venti.toml new file mode 100644 index 00000000000..557914adc6a --- /dev/null +++ b/providers/novita-ai/models/mindai/macaron-v1-venti.toml @@ -0,0 +1,13 @@ +# https://novita.ai/models/model-detail/mindai-macaron-v1-venti +# Effort: reasoning_effort = none|high|max; none suppresses reasoning. Graded baseline follows GLM-5.2/Macaron's native high|max modes; accepted aliases are not distinct levels. +# https://recipes.vllm.ai/mindlab-research/Macaron-V1-Coding-Venti +base_model = "mindlab/macaron-v1-venti" +reasoning_options = [{ type = "effort", values = ["none", "high", "max"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.825 +output = 2.475 +cache_read = 0.165 diff --git a/providers/novita-ai/models/minimax/minimax-m2.1.toml b/providers/novita-ai/models/minimax/minimax-m2.1.toml deleted file mode 100644 index a634f3db757..00000000000 --- a/providers/novita-ai/models/minimax/minimax-m2.1.toml +++ /dev/null @@ -1,27 +0,0 @@ -name = "Minimax M2.1" -description = "MiniMax model for chat, coding, office work, and agentic tasks" -family = "minimax" -release_date = "2025-12-23" -last_updated = "2025-12-23" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.3 -output = 1.2 -cache_read = 0.03 - -[limit] -context = 204_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/minimax/minimax-m2.5-highspeed.toml b/providers/novita-ai/models/minimax/minimax-m2.5-highspeed.toml deleted file mode 100644 index 8c7b8fdf5d5..00000000000 --- a/providers/novita-ai/models/minimax/minimax-m2.5-highspeed.toml +++ /dev/null @@ -1,28 +0,0 @@ -name = "MiniMax M2.5 Highspeed" -description = "High-speed MiniMax model for low-latency coding and agent workflows" -family = "minimax-m2.5" -release_date = "2026-02-12" -last_updated = "2026-02-12" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true -structured_output = true -open_weights = false - -[cost] -input = 0.6 -output = 2.4 -cache_read = 0.03 - -[limit] -context = 204_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/minimax/minimax-m2.5.toml b/providers/novita-ai/models/minimax/minimax-m2.5.toml index 3e39c3f9b77..c4b6b440c72 100644 --- a/providers/novita-ai/models/minimax/minimax-m2.5.toml +++ b/providers/novita-ai/models/minimax/minimax-m2.5.toml @@ -1,15 +1,11 @@ -name = "MiniMax M2.5" -description = "MiniMax model for chat, coding, office work, and agentic tasks" -family = "minimax" -release_date = "2026-02-12" -last_updated = "2026-02-12" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/minimax-minimax-m2.5 +# Always-on reasoning: thinking.type, enable_thinking, chat_template_kwargs.enable_thinking, and effort=none did not suppress it. +base_model = "minimax/MiniMax-M2.5" structured_output = true -open_weights = false +reasoning_options = [] + +[interleaved] +field = "reasoning_content" [cost] input = 0.3 @@ -17,12 +13,4 @@ output = 1.2 cache_read = 0.03 [limit] -context = 204_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] +output = 131_100 diff --git a/providers/novita-ai/models/minimax/minimax-m2.7-highspeed.toml b/providers/novita-ai/models/minimax/minimax-m2.7-highspeed.toml index 651f5b87863..d23af2e2cfa 100644 --- a/providers/novita-ai/models/minimax/minimax-m2.7-highspeed.toml +++ b/providers/novita-ai/models/minimax/minimax-m2.7-highspeed.toml @@ -1,8 +1,14 @@ +# Route is absent from Novita's public catalog but still answers live requests (2026-10-09). +# Always-on reasoning; enable_thinking, chat_template_kwargs.enable_thinking, and effort=none did not suppress it. +# Preserve curated pricing: /models does not return this route. base_model = "minimax/MiniMax-M2.7-highspeed" reasoning_options = [] last_updated = "2026-05-27" structured_output = true +[interleaved] +field = "reasoning_content" + [cost] input = 0.6 output = 2.4 diff --git a/providers/novita-ai/models/minimax/minimax-m2.7.toml b/providers/novita-ai/models/minimax/minimax-m2.7.toml index daaf0eb8fd3..c042c309ca3 100644 --- a/providers/novita-ai/models/minimax/minimax-m2.7.toml +++ b/providers/novita-ai/models/minimax/minimax-m2.7.toml @@ -1,28 +1,13 @@ -name = "MiniMax M2.7" -description = "MiniMax model for chat, coding, office work, and agentic tasks" -family = "minimax-m2.7" -release_date = "2026-03-18" -last_updated = "2026-03-18" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/minimax-minimax-m2.7 +# Always-on reasoning: thinking.type, enable_thinking, chat_template_kwargs.enable_thinking, and effort=none did not suppress it. +base_model = "minimax/MiniMax-M2.7" structured_output = true -open_weights = true +reasoning_options = [] + +[interleaved] +field = "reasoning_content" [cost] input = 0.3 output = 1.2 cache_read = 0.06 - -[limit] -context = 204_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/minimax/minimax-m2.toml b/providers/novita-ai/models/minimax/minimax-m2.toml deleted file mode 100644 index 108f84b5ce4..00000000000 --- a/providers/novita-ai/models/minimax/minimax-m2.toml +++ /dev/null @@ -1,27 +0,0 @@ -name = "MiniMax-M2" -description = "MiniMax model for chat, coding, office work, and agentic tasks" -family = "minimax" -release_date = "2025-10-27" -last_updated = "2025-10-27" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true -open_weights = true - -[cost] -input = 0.3 -output = 1.2 -cache_read = 0.03 - -[limit] -context = 204_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/minimax/minimax-m3.toml b/providers/novita-ai/models/minimax/minimax-m3.toml new file mode 100644 index 00000000000..ca05b6a19eb --- /dev/null +++ b/providers/novita-ai/models/minimax/minimax-m3.toml @@ -0,0 +1,24 @@ +# https://novita.ai/models/model-detail/minimax-minimax-m3 +# Toggle: thinking.type = enabled|disabled +# The second public pricing band starts at 524288 context tokens. +base_model = "minimax/MiniMax-M3" +structured_output = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.3 +output = 1.2 +cache_read = 0.06 + +[[cost.tiers]] +tier = { type = "context", size = 524_288 } +input = 0.6 +output = 2.4 +cache_read = 0.12 + +[limit] +context = 1_000_000 +output = 131_072 diff --git a/providers/novita-ai/models/minimaxai/minimax-m1-80k.toml b/providers/novita-ai/models/minimaxai/minimax-m1-80k.toml index 3188b27c055..90a1171d8be 100644 --- a/providers/novita-ai/models/minimaxai/minimax-m1-80k.toml +++ b/providers/novita-ai/models/minimaxai/minimax-m1-80k.toml @@ -1,23 +1,12 @@ -name = "MiniMax M1" -description = "MiniMax model for chat, coding, office work, and agentic tasks" -family = "minimax" -release_date = "2025-06-17" -last_updated = "2025-06-17" -attachment = false -reasoning = true +# https://novita.ai/models/model-detail/minimaxai-minimax-m1-80k +# Always-on reasoning is emitted as in ordinary content on this route. +base_model = "minimax/minimax-m1-80k" +structured_output = true reasoning_options = [] -temperature = true -tool_call = true -open_weights = true [cost] input = 0.55 output = 2.2 [limit] -context = 1_000_000 output = 40_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/mistralai/mistral-nemo.toml b/providers/novita-ai/models/mistralai/mistral-nemo.toml deleted file mode 100644 index 558a70af9a7..00000000000 --- a/providers/novita-ai/models/mistralai/mistral-nemo.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "Mistral Nemo" -description = "Mistral model for multilingual chat, reasoning, and tool-assisted workflows" -family = "mistral-nemo" -release_date = "2024-07-30" -last_updated = "2024-07-30" -attachment = false -reasoning = false -temperature = true -tool_call = false -structured_output = true -open_weights = true - -[cost] -input = 0.04 -output = 0.17 - -[limit] -context = 60_288 -output = 16_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/moonshotai/kimi-k2-0905.toml b/providers/novita-ai/models/moonshotai/kimi-k2-0905.toml index 4d4d36a3d63..3b4ad3dcbaa 100644 --- a/providers/novita-ai/models/moonshotai/kimi-k2-0905.toml +++ b/providers/novita-ai/models/moonshotai/kimi-k2-0905.toml @@ -1,24 +1,7 @@ -name = "Kimi K2 0905" -description = "Kimi model for long-context chat, coding, and agentic reasoning" -family = "kimi-k2" -release_date = "2025-09-05" -last_updated = "2025-09-05" -attachment = false -reasoning = false -temperature = true -knowledge = "2024-10" -tool_call = true +# https://novita.ai/models/model-detail/moonshotai-kimi-k2-0905 +base_model = "moonshotai/kimi-k2-0905" structured_output = true -open_weights = true [cost] input = 0.6 output = 2.5 - -[limit] -context = 262_144 -output = 98_304 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/moonshotai/kimi-k2-instruct.toml b/providers/novita-ai/models/moonshotai/kimi-k2-instruct.toml index 12b184ebf8e..9ae30b775a3 100644 --- a/providers/novita-ai/models/moonshotai/kimi-k2-instruct.toml +++ b/providers/novita-ai/models/moonshotai/kimi-k2-instruct.toml @@ -1,21 +1,9 @@ -name = "Kimi K2 Instruct" -description = "Kimi model for long-context chat, coding, and agentic reasoning" -release_date = "2025-07-11" -last_updated = "2025-07-11" -attachment = false -reasoning = false -temperature = true -tool_call = true -open_weights = true +# https://novita.ai/models/model-detail/moonshotai-kimi-k2-instruct +base_model = "moonshotai/kimi-k2-instruct" [cost] input = 0.57 output = 2.3 [limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] +output = 98_304 diff --git a/providers/novita-ai/models/moonshotai/kimi-k2-thinking.toml b/providers/novita-ai/models/moonshotai/kimi-k2-thinking.toml index 2deb70720c8..42dc7976797 100644 --- a/providers/novita-ai/models/moonshotai/kimi-k2-thinking.toml +++ b/providers/novita-ai/models/moonshotai/kimi-k2-thinking.toml @@ -1,15 +1,11 @@ -name = "Kimi K2 Thinking" -description = "Kimi reasoning model for long-horizon research, planning, and tool use" -family = "kimi-thinking" -release_date = "2025-11-07" -last_updated = "2026-06-29" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/moonshotai-kimi-k2-thinking +# Always-on reasoning; thinking.type=disabled is rejected. +base_model = "moonshotai/kimi-k2-thinking" structured_output = true -open_weights = true +reasoning_options = [] + +[interleaved] +field = "reasoning_content" [cost] input = 0.6 @@ -17,12 +13,4 @@ output = 2.5 cache_read = 0.15 [limit] -context = 262_144 output = 98_304 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/moonshotai/kimi-k2.5.toml b/providers/novita-ai/models/moonshotai/kimi-k2.5.toml index 2596e658e6f..88d11e8238f 100644 --- a/providers/novita-ai/models/moonshotai/kimi-k2.5.toml +++ b/providers/novita-ai/models/moonshotai/kimi-k2.5.toml @@ -1,29 +1,12 @@ -name = "Kimi K2.5" -description = "Kimi multimodal agent model for visual understanding, coding, and planning" -family = "kimi-k2" -release_date = "2026-01-27" -last_updated = "2026-01-27" -attachment = true -reasoning = true +# https://novita.ai/models/model-detail/moonshotai-kimi-k2.5 +# Toggle: thinking.type = enabled|disabled +base_model = "moonshotai/kimi-k2.5" reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -knowledge = "2025-01" -open_weights = true - -[cost] -input = 0.6 -output = 3.0 -cache_read = 0.1 - -[limit] -context = 262_144 -output = 262_144 [interleaved] field = "reasoning_content" -[modalities] -input = ["text", "image", "video"] -output = ["text"] +[cost] +input = 0.6 +output = 3 +cache_read = 0.1 diff --git a/providers/novita-ai/models/moonshotai/kimi-k2.6.toml b/providers/novita-ai/models/moonshotai/kimi-k2.6.toml index 37df0417686..a61098b78ee 100644 --- a/providers/novita-ai/models/moonshotai/kimi-k2.6.toml +++ b/providers/novita-ai/models/moonshotai/kimi-k2.6.toml @@ -1,3 +1,5 @@ +# https://novita.ai/models/model-detail/moonshotai-kimi-k2.6 +# Toggle: thinking.type = enabled|disabled base_model = "moonshotai/kimi-k2.6" reasoning_options = [{ type = "toggle" }] diff --git a/providers/novita-ai/models/moonshotai/kimi-k2.7-code.toml b/providers/novita-ai/models/moonshotai/kimi-k2.7-code.toml index e420a46934a..81c623b0f01 100644 --- a/providers/novita-ai/models/moonshotai/kimi-k2.7-code.toml +++ b/providers/novita-ai/models/moonshotai/kimi-k2.7-code.toml @@ -1,16 +1,12 @@ +# https://novita.ai/models/model-detail/moonshotai-kimi-k2.7-code +# Always-on reasoning (native baseline); thinking.type=disabled and effort=none did not suppress it. base_model = "moonshotai/kimi-k2.7-code" +reasoning_options = [] -[[reasoning_options]] -type = "toggle" +[interleaved] +field = "reasoning_content" [cost] input = 0.95 -output = 4.00 +output = 4 cache_read = 0.19 - -[limit] -context = 262_144 -output = 262_144 - -[modalities] -input = ["text", "image", "video"] diff --git a/providers/novita-ai/models/moonshotai/kimi-k3.toml b/providers/novita-ai/models/moonshotai/kimi-k3.toml index f398ea072a2..05e2e81b493 100644 --- a/providers/novita-ai/models/moonshotai/kimi-k3.toml +++ b/providers/novita-ai/models/moonshotai/kimi-k3.toml @@ -1,20 +1,15 @@ +# https://novita.ai/models/model-detail/moonshotai-kimi-k3 +# Effort: reasoning_effort = none|low|high|max; none suppresses reasoning on Novita, unlike the native Kimi API. base_model = "moonshotai/kimi-k3" +reasoning_options = [{ type = "effort", values = ["none", "low", "high", "max"] }] -[[reasoning_options]] -type = "toggle" - -[[reasoning_options]] -type = "effort" -values = ["low", "high", "max"] +[interleaved] +field = "reasoning_content" [cost] -input = 3.00 -output = 15.00 +input = 3 +output = 15 cache_read = 0.3 [limit] -context = 1_048_576 output = 1_048_576 - -[modalities] -input = ["text", "image", "video"] diff --git a/providers/novita-ai/models/nousresearch/hermes-2-pro-llama-3-8b.toml b/providers/novita-ai/models/nousresearch/hermes-2-pro-llama-3-8b.toml index 12e85fe19e7..2d2dfa94f11 100644 --- a/providers/novita-ai/models/nousresearch/hermes-2-pro-llama-3-8b.toml +++ b/providers/novita-ai/models/nousresearch/hermes-2-pro-llama-3-8b.toml @@ -1,22 +1,6 @@ -name = "Hermes 2 Pro Llama 3 8B" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -release_date = "2024-06-27" -last_updated = "2024-06-27" -attachment = false -reasoning = false -temperature = true -tool_call = false -structured_output = true -open_weights = true +# https://novita.ai/models/model-detail/nousresearch-hermes-2-pro-llama-3-8b +base_model = "nousresearch/hermes-2-pro-llama-3-8b" [cost] input = 0.14 output = 0.14 - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/nvidia/nemotron-3-nano-30b-a3b.toml b/providers/novita-ai/models/nvidia/nemotron-3-nano-30b-a3b.toml new file mode 100644 index 00000000000..34cfcdbbd3e --- /dev/null +++ b/providers/novita-ai/models/nvidia/nemotron-3-nano-30b-a3b.toml @@ -0,0 +1,15 @@ +# https://novita.ai/models/model-detail/nvidia-nemotron-3-nano-30b-a3b +# Toggle: thinking.type = enabled|disabled +base_model = "nvidia/nemotron-3-nano-30b-a3b" +structured_output = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.05 +output = 0.2 + +[limit] +output = 32_768 diff --git a/providers/novita-ai/models/openai/gpt-oss-120b.toml b/providers/novita-ai/models/openai/gpt-oss-120b.toml index f13f4b73585..c782c133889 100644 --- a/providers/novita-ai/models/openai/gpt-oss-120b.toml +++ b/providers/novita-ai/models/openai/gpt-oss-120b.toml @@ -1,23 +1,12 @@ -name = "OpenAI GPT OSS 120B" -description = "Open-weight GPT model for self-hosted reasoning and instruction-following workloads" -release_date = "2025-08-06" -last_updated = "2025-08-06" -attachment = true -reasoning = true +# https://novita.ai/models/model-detail/openai-gpt-oss-120b +# Effort: reasoning_effort = low|medium|high (native baseline). thinking.type=disabled and effort=none do not suppress reasoning. +# The inventory's image-input label is not supported by the underlying text-only model. +base_model = "openai/gpt-oss-120b" reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }] -temperature = true -tool_call = true -structured_output = true -open_weights = true + +[interleaved] +field = "reasoning_content" [cost] input = 0.05 output = 0.25 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/openai/gpt-oss-20b.toml b/providers/novita-ai/models/openai/gpt-oss-20b.toml deleted file mode 100644 index 0fdae660ba1..00000000000 --- a/providers/novita-ai/models/openai/gpt-oss-20b.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "OpenAI: GPT OSS 20B" -description = "Open-weight GPT model for self-hosted reasoning and instruction-following workloads" -release_date = "2025-08-06" -last_updated = "2025-08-06" -attachment = true -reasoning = true -reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }] -temperature = true -tool_call = false -structured_output = true -open_weights = true - -[cost] -input = 0.04 -output = 0.15 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/paddlepaddle/paddleocr-vl.toml b/providers/novita-ai/models/paddlepaddle/paddleocr-vl.toml index 6dfce843e12..94c15d21c84 100644 --- a/providers/novita-ai/models/paddlepaddle/paddleocr-vl.toml +++ b/providers/novita-ai/models/paddlepaddle/paddleocr-vl.toml @@ -1,12 +1,5 @@ -name = "PaddleOCR-VL" -description = "Multimodal model for analyzing text, images, documents, and rich media" -release_date = "2025-10-22" -last_updated = "2025-10-22" -attachment = true -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/paddlepaddle-paddleocr-vl +base_model = "paddlepaddle/paddleocr-vl" [cost] input = 0.02 @@ -14,8 +7,3 @@ output = 0.02 [limit] context = 16_384 -output = 16_384 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen-2.5-72b-instruct.toml b/providers/novita-ai/models/qwen/qwen-2.5-72b-instruct.toml deleted file mode 100644 index 6dfd1b9d503..00000000000 --- a/providers/novita-ai/models/qwen/qwen-2.5-72b-instruct.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "Qwen 2.5 72B Instruct" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -family = "qwen" -release_date = "2024-10-15" -last_updated = "2024-10-15" -attachment = false -reasoning = false -temperature = true -knowledge = "2024-04" -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.38 -output = 0.4 - -[limit] -context = 32_000 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen-mt-plus.toml b/providers/novita-ai/models/qwen/qwen-mt-plus.toml index 8591838a922..5dd2bb3b81d 100644 --- a/providers/novita-ai/models/qwen/qwen-mt-plus.toml +++ b/providers/novita-ai/models/qwen/qwen-mt-plus.toml @@ -1,21 +1,6 @@ -name = "Qwen MT Plus" -description = "Translation model for multilingual conversion, localization, and cross-language workflows" -release_date = "2025-09-03" -last_updated = "2025-09-03" -attachment = false -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/qwen-qwen-mt-plus +base_model = "alibaba/qwen-mt-plus" [cost] input = 0.25 output = 0.75 - -[limit] -context = 16_384 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen2.5-7b-instruct.toml b/providers/novita-ai/models/qwen/qwen2.5-7b-instruct.toml deleted file mode 100644 index a7e408e56f0..00000000000 --- a/providers/novita-ai/models/qwen/qwen2.5-7b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen2.5 7B Instruct" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -release_date = "2025-04-16" -last_updated = "2025-04-16" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.07 -output = 0.07 - -[limit] -context = 32_000 -output = 32_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen2.5-vl-72b-instruct.toml b/providers/novita-ai/models/qwen/qwen2.5-vl-72b-instruct.toml deleted file mode 100644 index c48fdbc92cb..00000000000 --- a/providers/novita-ai/models/qwen/qwen2.5-vl-72b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen2.5 VL 72B Instruct" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -family = "qwen" -release_date = "2025-03-25" -last_updated = "2025-03-25" -attachment = true -reasoning = false -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.8 -output = 0.8 - -[limit] -context = 32_768 -output = 32_768 - -[modalities] -input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-235b-a22b-fp8.toml b/providers/novita-ai/models/qwen/qwen3-235b-a22b-fp8.toml deleted file mode 100644 index 5794dec5f57..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-235b-a22b-fp8.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 235B A22B" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -release_date = "2025-04-29" -last_updated = "2025-04-29" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.2 -output = 0.8 - -[limit] -context = 40_960 -output = 20_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-235b-a22b-instruct-2507.toml b/providers/novita-ai/models/qwen/qwen3-235b-a22b-instruct-2507.toml deleted file mode 100644 index a260dd0396f..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-235b-a22b-instruct-2507.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "Qwen3 235B A22B Instruct 2507" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -family = "qwen" -release_date = "2025-07-22" -last_updated = "2025-07-22" -attachment = false -reasoning = false -temperature = true -knowledge = "2025-04" -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.09 -output = 0.58 - -[limit] -context = 131_072 -output = 16_384 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-235b-a22b-thinking-2507.toml b/providers/novita-ai/models/qwen/qwen3-235b-a22b-thinking-2507.toml deleted file mode 100644 index bf9b1f12b4a..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-235b-a22b-thinking-2507.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "Qwen3 235B A22b Thinking 2507" -description = "Qwen reasoning model for deliberate problem solving, math, and coding" -family = "qwen" -release_date = "2025-07-25" -last_updated = "2025-07-25" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -knowledge = "2025-04" -tool_call = true -open_weights = true - -[cost] -input = 0.3 -output = 3.0 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-30b-a3b-fp8.toml b/providers/novita-ai/models/qwen/qwen3-30b-a3b-fp8.toml deleted file mode 100644 index 04d55d6b9a8..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-30b-a3b-fp8.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 30B A3B" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -release_date = "2025-04-29" -last_updated = "2025-04-29" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.09 -output = 0.45 - -[limit] -context = 40_960 -output = 20_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-32b-fp8.toml b/providers/novita-ai/models/qwen/qwen3-32b-fp8.toml deleted file mode 100644 index cf31ae85636..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-32b-fp8.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 32B" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -release_date = "2025-04-29" -last_updated = "2025-04-29" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.1 -output = 0.45 - -[limit] -context = 40_960 -output = 20_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-4b-fp8.toml b/providers/novita-ai/models/qwen/qwen3-4b-fp8.toml deleted file mode 100644 index 939cd8cee7f..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-4b-fp8.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 4B" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -release_date = "2025-04-29" -last_updated = "2025-04-29" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.03 -output = 0.03 - -[limit] -context = 128_000 -output = 20_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-8b-fp8.toml b/providers/novita-ai/models/qwen/qwen3-8b-fp8.toml deleted file mode 100644 index 4b5c5d01dec..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-8b-fp8.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 8B" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -release_date = "2025-04-29" -last_updated = "2025-04-29" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.035 -output = 0.138 - -[limit] -context = 128_000 -output = 20_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-coder-30b-a3b-instruct.toml b/providers/novita-ai/models/qwen/qwen3-coder-30b-a3b-instruct.toml deleted file mode 100644 index 69602c25a80..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-coder-30b-a3b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 Coder 30b A3B Instruct" -description = "Qwen coding model for software agents, repository edits, and code reasoning" -release_date = "2025-10-09" -last_updated = "2025-10-09" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.07 -output = 0.27 - -[limit] -context = 160_000 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-coder-480b-a35b-instruct.toml b/providers/novita-ai/models/qwen/qwen3-coder-480b-a35b-instruct.toml deleted file mode 100644 index c45aca69cae..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-coder-480b-a35b-instruct.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "Qwen3 Coder 480B A35B Instruct" -description = "Qwen coding model for software agents, repository edits, and code reasoning" -family = "qwen" -release_date = "2025-07-23" -last_updated = "2025-07-23" -attachment = false -reasoning = false -temperature = true -knowledge = "2025-04" -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.38 -output = 1.55 - -[limit] -context = 262_144 -output = 65_536 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-coder-next.toml b/providers/novita-ai/models/qwen/qwen3-coder-next.toml deleted file mode 100644 index 9ad4ec23a79..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-coder-next.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "Qwen3 Coder Next" -description = "Qwen coding model for software agents, repository edits, and code reasoning" -family = "qwen" -release_date = "2026-02-03" -last_updated = "2026-02-03" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.2 -output = 1.5 - -[limit] -context = 262_144 -output = 65_536 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-embedding-8b.toml b/providers/novita-ai/models/qwen/qwen3-embedding-8b.toml new file mode 100644 index 00000000000..6996e37241d --- /dev/null +++ b/providers/novita-ai/models/qwen/qwen3-embedding-8b.toml @@ -0,0 +1,7 @@ +# https://novita.ai/models/model-detail/qwen-qwen3-embedding-8b +# POST /openai/v1/embeddings returns a 4096-dimensional vector; output is not a generative token limit. +base_model = "alibaba/qwen3-embedding-8b" + +[cost] +input = 0.07 +output = 0 diff --git a/providers/novita-ai/models/qwen/qwen3-max.toml b/providers/novita-ai/models/qwen/qwen3-max.toml deleted file mode 100644 index c25a772eb91..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-max.toml +++ /dev/null @@ -1,24 +0,0 @@ -name = "Qwen3 Max" -description = "Flagship Qwen model for complex reasoning, coding, and agentic workflows" -family = "qwen" -release_date = "2025-09-24" -last_updated = "2025-09-24" -attachment = false -reasoning = false -temperature = true -knowledge = "2025-04" -tool_call = true -structured_output = true -open_weights = false - -[cost] -input = 2.11 -output = 8.45 - -[limit] -context = 262_144 -output = 65_536 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-next-80b-a3b-instruct.toml b/providers/novita-ai/models/qwen/qwen3-next-80b-a3b-instruct.toml deleted file mode 100644 index 554db812d4c..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-next-80b-a3b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 Next 80B A3B Instruct" -description = "Qwen instruction model for multilingual chat, reasoning, and tool use" -release_date = "2025-09-10" -last_updated = "2025-09-10" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.15 -output = 1.5 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-next-80b-a3b-thinking.toml b/providers/novita-ai/models/qwen/qwen3-next-80b-a3b-thinking.toml deleted file mode 100644 index 873ac60137b..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-next-80b-a3b-thinking.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "Qwen3 Next 80B A3B Thinking" -description = "Qwen reasoning model for deliberate problem solving, math, and coding" -release_date = "2025-09-10" -last_updated = "2025-09-10" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.15 -output = 1.5 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-instruct.toml b/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-instruct.toml deleted file mode 100644 index 22f2b93a22e..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-instruct.toml +++ /dev/null @@ -1,25 +0,0 @@ -name = "Qwen3 Omni 30B A3B Instruct" -description = "Qwen omni model for text, vision, audio, and multimodal agent tasks" -family = "qwen" -release_date = "2025-09-24" -last_updated = "2025-09-24" -attachment = true -reasoning = false -temperature = true -knowledge = "2024-04" -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.25 -output = 0.97 -input_audio = 2.2 # Cost per million audio input tokens (USD) -output_audio = 1.788 -[limit] -context = 65_536 -output = 16_384 - -[modalities] -input = ["text", "video", "audio", "image"] -output = ["text", "audio"] diff --git a/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml b/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml index 36091ee4f16..b2c88e0adf9 100644 --- a/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml +++ b/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml @@ -1,25 +1,12 @@ -name = "Qwen3 Omni 30B A3B Thinking" -description = "Qwen omni model for text, vision, audio, and multimodal agent tasks" -release_date = "2025-09-24" -last_updated = "2025-09-24" -attachment = true -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/qwen-qwen3-omni-30b-a3b-thinking +# Live requests emit no reasoning by default or when disabled; thinking.type=enabled and enable_thinking=true are rejected. +# This route currently does not expose the underlying checkpoint's thinking mode. +# Preserve the curated input-audio rate, which the inventory does not return; output-audio pricing is inapplicable to this text-output checkpoint. +base_model = "alibaba/qwen3-omni-30b-a3b-thinking" +reasoning = false structured_output = true -open_weights = true [cost] input = 0.25 output = 0.97 -input_audio = 2.2 # Cost per million audio input tokens (USD) -output_audio = 1.788 - -[limit] -context = 65_536 -output = 16_384 - -[modalities] -input = ["text", "audio", "video", "image"] -output = ["text"] +input_audio = 2.2 diff --git a/providers/novita-ai/models/qwen/qwen3-reranker-8b.toml b/providers/novita-ai/models/qwen/qwen3-reranker-8b.toml new file mode 100644 index 00000000000..355f43757fd --- /dev/null +++ b/providers/novita-ai/models/qwen/qwen3-reranker-8b.toml @@ -0,0 +1,8 @@ +# https://novita.ai/models/model-detail/qwen-qwen3-reranker-8b +# https://novita.ai/docs/api-reference/model-apis-llm-create-rerank +# Public description identifies the text Qwen3 Reranker 8B; the site's HF mirror URL points to a different VL checkpoint. +base_model = "alibaba/qwen3-reranker-8b" + +[cost] +input = 0.05 +output = 0.05 diff --git a/providers/novita-ai/models/qwen/qwen3-vl-235b-a22b-instruct.toml b/providers/novita-ai/models/qwen/qwen3-vl-235b-a22b-instruct.toml deleted file mode 100644 index 5a7c4c6f0f9..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-vl-235b-a22b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 VL 235B A22B Instruct" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -release_date = "2025-09-24" -last_updated = "2025-09-24" -attachment = true -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.3 -output = 1.5 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-vl-235b-a22b-thinking.toml b/providers/novita-ai/models/qwen/qwen3-vl-235b-a22b-thinking.toml deleted file mode 100644 index d3b458466c6..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-vl-235b-a22b-thinking.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Qwen3 VL 235B A22B Thinking" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -release_date = "2025-09-24" -last_updated = "2025-09-24" -attachment = true -reasoning = true -reasoning_options = [] -temperature = true -tool_call = false -open_weights = true - -[cost] -input = 0.98 -output = 3.95 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-vl-30b-a3b-instruct.toml b/providers/novita-ai/models/qwen/qwen3-vl-30b-a3b-instruct.toml index 92b9d517e9d..95c1c59045e 100644 --- a/providers/novita-ai/models/qwen/qwen3-vl-30b-a3b-instruct.toml +++ b/providers/novita-ai/models/qwen/qwen3-vl-30b-a3b-instruct.toml @@ -1,13 +1,6 @@ -name = "qwen/qwen3-vl-30b-a3b-instruct" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -release_date = "2025-10-11" -last_updated = "2025-10-11" -attachment = true -reasoning = false -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/qwen-qwen3-vl-30b-a3b-instruct +base_model = "alibaba/qwen3-vl-30b-a3b-instruct" structured_output = true -open_weights = true [cost] input = 0.2 @@ -15,8 +8,3 @@ output = 0.7 [limit] context = 131_072 -output = 32_768 - -[modalities] -input = ["text", "video", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-vl-30b-a3b-thinking.toml b/providers/novita-ai/models/qwen/qwen3-vl-30b-a3b-thinking.toml deleted file mode 100644 index 1d82f3168cf..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-vl-30b-a3b-thinking.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "qwen/qwen3-vl-30b-a3b-thinking" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -release_date = "2025-10-11" -last_updated = "2025-10-11" -attachment = true -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.2 -output = 1.0 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3-vl-8b-instruct.toml b/providers/novita-ai/models/qwen/qwen3-vl-8b-instruct.toml deleted file mode 100644 index b2855bb05e3..00000000000 --- a/providers/novita-ai/models/qwen/qwen3-vl-8b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "qwen/qwen3-vl-8b-instruct" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -release_date = "2025-10-17" -last_updated = "2025-10-17" -attachment = true -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.08 -output = 0.5 - -[limit] -context = 131_072 -output = 32_768 - -[modalities] -input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3.5-122b-a10b.toml b/providers/novita-ai/models/qwen/qwen3.5-122b-a10b.toml index e8529355c39..685e84efb7d 100644 --- a/providers/novita-ai/models/qwen/qwen3.5-122b-a10b.toml +++ b/providers/novita-ai/models/qwen/qwen3.5-122b-a10b.toml @@ -1,24 +1,15 @@ -name = "Qwen3.5-122B-A10B" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -family = "qwen" -release_date = "2026-02-26" -last_updated = "2026-02-26" -attachment = true -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -open_weights = true +# https://novita.ai/models/model-detail/qwen-qwen3.5-122b-a10b +# Toggle: thinking.type = enabled|disabled +# Budget: thinking_budget (reasoning tokens) +base_model = "alibaba/qwen3.5-122b-a10b" +reasoning_options = [{ type = "toggle" }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.4 output = 3.2 -[limit] -context = 262_144 -output = 65_536 - [modalities] input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3.5-27b.toml b/providers/novita-ai/models/qwen/qwen3.5-27b.toml index 370ab051428..96e0a29e0de 100644 --- a/providers/novita-ai/models/qwen/qwen3.5-27b.toml +++ b/providers/novita-ai/models/qwen/qwen3.5-27b.toml @@ -1,24 +1,15 @@ -name = "Qwen3.5-27B" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -family = "qwen" -release_date = "2026-02-26" -last_updated = "2026-02-26" -attachment = true -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -open_weights = true +# https://novita.ai/models/model-detail/qwen-qwen3.5-27b +# Toggle: thinking.type = enabled|disabled +# Budget: thinking_budget (reasoning tokens) +base_model = "alibaba/qwen3.5-27b" +reasoning_options = [{ type = "toggle" }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.3 output = 2.4 -[limit] -context = 262_144 -output = 65_536 - [modalities] input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3.5-35b-a3b.toml b/providers/novita-ai/models/qwen/qwen3.5-35b-a3b.toml index a691cc1925c..d959ca015d4 100644 --- a/providers/novita-ai/models/qwen/qwen3.5-35b-a3b.toml +++ b/providers/novita-ai/models/qwen/qwen3.5-35b-a3b.toml @@ -1,24 +1,15 @@ -name = "Qwen3.5-35B-A3B" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -family = "qwen" -release_date = "2026-02-26" -last_updated = "2026-02-26" -attachment = true -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -open_weights = true +# https://novita.ai/models/model-detail/qwen-qwen3.5-35b-a3b +# Toggle: thinking.type = enabled|disabled +# Budget: thinking_budget (reasoning tokens) +base_model = "alibaba/qwen3.5-35b-a3b" +reasoning_options = [{ type = "toggle" }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.25 -output = 2.0 - -[limit] -context = 262_144 -output = 65_536 +output = 2 [modalities] input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3.5-397b-a17b.toml b/providers/novita-ai/models/qwen/qwen3.5-397b-a17b.toml index 43800444813..ba0068811bb 100644 --- a/providers/novita-ai/models/qwen/qwen3.5-397b-a17b.toml +++ b/providers/novita-ai/models/qwen/qwen3.5-397b-a17b.toml @@ -1,24 +1,15 @@ -name = "Qwen3.5-397B-A17B" -description = "Qwen vision-language model for visual reasoning, documents, and agent tasks" -family = "qwen" -release_date = "2026-02-17" -last_updated = "2026-02-17" -attachment = true -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -open_weights = true +# https://novita.ai/models/model-detail/qwen-qwen3.5-397b-a17b +# Toggle: thinking.type = enabled|disabled +# Budget: thinking_budget (reasoning tokens) +base_model = "alibaba/qwen3.5-397b-a17b" +reasoning_options = [{ type = "toggle" }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.6 output = 3.6 -[limit] -context = 262_144 -output = 64_000 - [modalities] input = ["text", "image", "video"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3.6-27b.toml b/providers/novita-ai/models/qwen/qwen3.6-27b.toml new file mode 100644 index 00000000000..f58ff0b139a --- /dev/null +++ b/providers/novita-ai/models/qwen/qwen3.6-27b.toml @@ -0,0 +1,15 @@ +# https://novita.ai/models/model-detail/qwen-qwen3.6-27b +# Toggle: thinking.type = enabled|disabled +# Budget: thinking_budget (reasoning tokens) +base_model = "alibaba/qwen3.6-27b" +reasoning_options = [{ type = "toggle" }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.6 +output = 3.6 + +[modalities] +input = ["text", "image", "video"] diff --git a/providers/novita-ai/models/qwen/qwen3.6-35b-a3b.toml b/providers/novita-ai/models/qwen/qwen3.6-35b-a3b.toml new file mode 100644 index 00000000000..20edf595665 --- /dev/null +++ b/providers/novita-ai/models/qwen/qwen3.6-35b-a3b.toml @@ -0,0 +1,15 @@ +# https://novita.ai/models/model-detail/qwen-qwen3.6-35b-a3b +# Toggle: thinking.type = enabled|disabled +# Budget: thinking_budget (reasoning tokens) +base_model = "alibaba/qwen3.6-35b-a3b" +reasoning_options = [{ type = "toggle" }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.248 +output = 1.485 + +[modalities] +input = ["text", "image", "video"] diff --git a/providers/novita-ai/models/qwen/qwen3.7-max.toml b/providers/novita-ai/models/qwen/qwen3.7-max.toml index 363a8ba3b78..b1a186eefa0 100644 --- a/providers/novita-ai/models/qwen/qwen3.7-max.toml +++ b/providers/novita-ai/models/qwen/qwen3.7-max.toml @@ -1,26 +1,15 @@ -name = "Qwen3.7-Max" -description = "Flagship Qwen model for complex reasoning, coding, and agentic workflows" -family = "qwen" -release_date = "2026-05-21" -last_updated = "2026-05-27" -attachment = false -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/qwen-qwen3.7-max +# Toggle: thinking.type = enabled|disabled +# Budget: thinking_budget (reasoning tokens) +# Public pricing does not advertise a cache-write charge. +base_model = "alibaba/qwen3.7-max" structured_output = true -open_weights = false +reasoning_options = [{ type = "toggle" }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" [cost] input = 1.25 output = 3.75 cache_read = 0.25 -cache_write = 1.5625 - -[limit] -context = 1_000_000 -output = 65_536 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/qwen/qwen3.8-2.4t-a95b.toml b/providers/novita-ai/models/qwen/qwen3.8-2.4t-a95b.toml new file mode 100644 index 00000000000..bc43cab84b1 --- /dev/null +++ b/providers/novita-ai/models/qwen/qwen3.8-2.4t-a95b.toml @@ -0,0 +1,16 @@ +# https://novita.ai/models/model-detail/qwen-qwen3.8-2.4t-a95b +# Effort: reasoning_effort = low|medium|xhigh; thinking.type=disabled is rejected. +# Budget: thinking_budget (reasoning tokens); cannot be combined with effort. +base_model = "alibaba/qwen3.8-2.4t-a95b" +reasoning_options = [{ type = "effort", values = ["low", "medium", "xhigh"] }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 2 +output = 6 +cache_read = 0.25 + +[limit] +context = 1_000_000 diff --git a/providers/novita-ai/models/qwen/qwen3.8-27b.toml b/providers/novita-ai/models/qwen/qwen3.8-27b.toml new file mode 100644 index 00000000000..009a208e36b --- /dev/null +++ b/providers/novita-ai/models/qwen/qwen3.8-27b.toml @@ -0,0 +1,18 @@ +# https://novita.ai/models/model-detail/qwen-qwen3.8-27b +# Toggle: thinking.type = enabled|disabled +# Effort: reasoning_effort = low|medium|xhigh +# Budget: thinking_budget (reasoning tokens); cannot be combined with effort. +base_model = "alibaba/qwen3.8-27b" +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "medium", "xhigh"] }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.42 +output = 3 +cache_read = 0.085 + +[limit] +context = 1_000_000 +output = 131_072 diff --git a/providers/novita-ai/models/qwen/qwen3.8-flash.toml b/providers/novita-ai/models/qwen/qwen3.8-flash.toml new file mode 100644 index 00000000000..f668634487e --- /dev/null +++ b/providers/novita-ai/models/qwen/qwen3.8-flash.toml @@ -0,0 +1,14 @@ +# https://novita.ai/models/model-detail/qwen-qwen3.8-flash +# Toggle: thinking.type = enabled|disabled +# Effort: reasoning_effort = low|medium|xhigh +# Budget: thinking_budget (reasoning tokens); cannot be combined with effort. +base_model = "alibaba/qwen3.8-flash" +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "medium", "xhigh"] }, { type = "budget_tokens" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.15 +output = 0.47 +cache_read = 0.016 diff --git a/providers/novita-ai/models/qwen/qwen3.8-max.toml b/providers/novita-ai/models/qwen/qwen3.8-max.toml new file mode 100644 index 00000000000..f7cbd838d05 --- /dev/null +++ b/providers/novita-ai/models/qwen/qwen3.8-max.toml @@ -0,0 +1,18 @@ +# https://novita.ai/models/model-detail/qwen-qwen3.8-max +# Toggle: thinking.type = enabled|disabled +# Effort: reasoning_effort = low|medium|xhigh +# thinking_budget requests are rejected on this route; do not inherit the lab API's budget control. +base_model = "alibaba/qwen3.8-max" +structured_output = true +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "medium", "xhigh"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 2 +output = 6 +cache_read = 0.25 + +[modalities] +input = ["text", "image", "video"] diff --git a/providers/novita-ai/models/sao10K/L3-8B-stheno-v3.2.toml b/providers/novita-ai/models/sao10K/L3-8B-stheno-v3.2.toml deleted file mode 100644 index 04efbfb5da6..00000000000 --- a/providers/novita-ai/models/sao10K/L3-8B-stheno-v3.2.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "L3 8B Stheno V3.2" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-11-29" -last_updated = "2024-11-29" -attachment = false -reasoning = false -temperature = true -tool_call = true -open_weights = true - -[cost] -input = 0.05 -output = 0.05 - -[limit] -context = 8_192 -output = 32_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/sao10K/l3-70b-euryale-v2.1.toml b/providers/novita-ai/models/sao10K/l3-70b-euryale-v2.1.toml deleted file mode 100644 index 900d618f62c..00000000000 --- a/providers/novita-ai/models/sao10K/l3-70b-euryale-v2.1.toml +++ /dev/null @@ -1,21 +0,0 @@ -name = "L3 70B Euryale V2.1 " -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -release_date = "2024-06-18" -last_updated = "2024-06-18" -attachment = false -reasoning = false -temperature = true -tool_call = true -open_weights = true - -[cost] -input = 1.48 -output = 1.48 - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/sao10K/l3-8b-lunaris.toml b/providers/novita-ai/models/sao10K/l3-8b-lunaris.toml deleted file mode 100644 index 57b6a8bf655..00000000000 --- a/providers/novita-ai/models/sao10K/l3-8b-lunaris.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Sao10k L3 8B Lunaris " -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -release_date = "2024-11-28" -last_updated = "2024-11-28" -attachment = false -reasoning = false -temperature = true -tool_call = false -structured_output = true -open_weights = true - -[cost] -input = 0.05 -output = 0.05 - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/sao10K/l31-70b-euryale-v2.2.toml b/providers/novita-ai/models/sao10K/l31-70b-euryale-v2.2.toml deleted file mode 100644 index a28c8d7cf82..00000000000 --- a/providers/novita-ai/models/sao10K/l31-70b-euryale-v2.2.toml +++ /dev/null @@ -1,21 +0,0 @@ -name = "L31 70B Euryale V2.2" -description = "Open-weight instruction model for adaptable chat and self-hosted production workloads" -release_date = "2024-09-19" -last_updated = "2024-09-19" -attachment = false -reasoning = false -temperature = true -tool_call = true -open_weights = true - -[cost] -input = 1.48 -output = 1.48 - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/sao10k/l3-70b-euryale-v2.1.toml b/providers/novita-ai/models/sao10k/l3-70b-euryale-v2.1.toml new file mode 100644 index 00000000000..899f57fb037 --- /dev/null +++ b/providers/novita-ai/models/sao10k/l3-70b-euryale-v2.1.toml @@ -0,0 +1,8 @@ +# https://novita.ai/models/model-detail/sao10k-l3-70b-euryale-v2.1 +# Preserve the exact case-sensitive model ID returned by Novita. +base_model = "sao10k/l3-70b-euryale-v2.1" +tool_call = true + +[cost] +input = 1.48 +output = 1.48 diff --git a/providers/novita-ai/models/stepfun/step-3.7-flash.toml b/providers/novita-ai/models/stepfun/step-3.7-flash.toml new file mode 100644 index 00000000000..947d71a8b47 --- /dev/null +++ b/providers/novita-ai/models/stepfun/step-3.7-flash.toml @@ -0,0 +1,16 @@ +# https://novita.ai/models/model-detail/stepfun-step-3.7-flash +# Effort: reasoning_effort = low|medium|high (native baseline); thinking.type=disabled does not suppress reasoning. +base_model = "stepfun/step-3.7-flash" +structured_output = true +reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.2 +output = 1.15 +cache_read = 0.04 + +[limit] +context = 262_144 diff --git a/providers/novita-ai/models/tencent/hy3.toml b/providers/novita-ai/models/tencent/hy3.toml new file mode 100644 index 00000000000..41773441988 --- /dev/null +++ b/providers/novita-ai/models/tencent/hy3.toml @@ -0,0 +1,17 @@ +# https://novita.ai/models/model-detail/tencent-hy3 +# Effort: reasoning_effort = none|high; none suppresses reasoning. Native effort baseline retained; accepted aliases are not additional distinct levels. +base_model = "tencent/hy3" +structured_output = true +reasoning_options = [{ type = "effort", values = ["none", "high"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.14 +output = 0.58 +cache_read = 0.035 + +[limit] +context = 262_144 +output = 262_144 diff --git a/providers/novita-ai/models/tencent/hy4-preview.toml b/providers/novita-ai/models/tencent/hy4-preview.toml new file mode 100644 index 00000000000..3a470763d85 --- /dev/null +++ b/providers/novita-ai/models/tencent/hy4-preview.toml @@ -0,0 +1,16 @@ +# https://novita.ai/models/model-detail/tencent-hy4-preview +# Effort: reasoning_effort = none|high; none suppresses reasoning. Native effort baseline retained; accepted aliases are not additional distinct levels. +base_model = "tencent/hy4-preview" +structured_output = true +reasoning_options = [{ type = "effort", values = ["none", "high"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.834 +output = 2.501 +cache_read = 0.042 + +[limit] +context = 1_000_000 diff --git a/providers/novita-ai/models/xiaomimimo/mimo-v2-flash.toml b/providers/novita-ai/models/xiaomimimo/mimo-v2-flash.toml deleted file mode 100644 index e7d77465208..00000000000 --- a/providers/novita-ai/models/xiaomimimo/mimo-v2-flash.toml +++ /dev/null @@ -1,26 +0,0 @@ -name = "XiaomiMiMo/MiMo-V2-Flash" -description = "MiMo flash model for fast multimodal assistance and agent workflows" -family = "mimo" -release_date = "2025-12-19" -last_updated = "2025-12-19" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -knowledge = "2024-12" -tool_call = true -structured_output = true -open_weights = true - -[cost] -input = 0.1 -output = 0.3 -cache_read = 0.30 - -[limit] -context = 262_144 -output = 32_000 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/xiaomimimo/mimo-v2-pro.toml b/providers/novita-ai/models/xiaomimimo/mimo-v2-pro.toml deleted file mode 100644 index 54890ef6778..00000000000 --- a/providers/novita-ai/models/xiaomimimo/mimo-v2-pro.toml +++ /dev/null @@ -1,18 +0,0 @@ -base_model = "xiaomi/mimo-v2-pro" -reasoning_options = [] -last_updated = "2026-05-27" -structured_output = true - -[interleaved] -field = "reasoning_content" - -[cost] -input = 2 -output = 6 -cache_read = 0.4 - -[[cost.tiers]] -tier = { type = "context", size = 256_000 } -input = 2 -output = 6 -cache_read = 0.4 diff --git a/providers/novita-ai/models/xiaomimimo/mimo-v2.5-pro.toml b/providers/novita-ai/models/xiaomimimo/mimo-v2.5-pro.toml index 5a921aebd83..6f179200b70 100644 --- a/providers/novita-ai/models/xiaomimimo/mimo-v2.5-pro.toml +++ b/providers/novita-ai/models/xiaomimimo/mimo-v2.5-pro.toml @@ -1,7 +1,9 @@ +# https://novita.ai/models/model-detail/xiaomimimo-mimo-v2.5-pro +# Toggle: thinking.type = enabled|disabled +# Public pricing is flat; the old 256K tier duplicated the base price. base_model = "xiaomi/mimo-v2.5-pro" -reasoning_options = [] -last_updated = "2026-05-27" structured_output = true +reasoning_options = [{ type = "toggle" }] [interleaved] field = "reasoning_content" @@ -10,9 +12,3 @@ field = "reasoning_content" input = 0.522 output = 1.044 cache_read = 0.0043 - -[[cost.tiers]] -tier = { type = "context", size = 256_000 } -input = 0.522 -output = 1.044 -cache_read = 0.0043 diff --git a/providers/novita-ai/models/xiaomimimo/mimo-v2.5.toml b/providers/novita-ai/models/xiaomimimo/mimo-v2.5.toml new file mode 100644 index 00000000000..e2fe08d7083 --- /dev/null +++ b/providers/novita-ai/models/xiaomimimo/mimo-v2.5.toml @@ -0,0 +1,16 @@ +# https://novita.ai/models/model-detail/xiaomimimo-mimo-v2.5 +# Toggle: thinking.type = enabled|disabled +base_model = "xiaomi/mimo-v2.5" +structured_output = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.168 +output = 0.336 +cache_read = 0.0034 + +[modalities] +input = ["text", "image", "video"] diff --git a/providers/novita-ai/models/xiaomimimo/mimo-v2.6-flash.toml b/providers/novita-ai/models/xiaomimimo/mimo-v2.6-flash.toml new file mode 100644 index 00000000000..9d7b5437030 --- /dev/null +++ b/providers/novita-ai/models/xiaomimimo/mimo-v2.6-flash.toml @@ -0,0 +1,13 @@ +# https://novita.ai/models/model-detail/xiaomimimo-mimo-v2.6-flash +# Toggle: thinking.type = enabled|disabled +base_model = "xiaomi/mimo-v2.6-flash" +structured_output = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.14 +output = 0.28 +cache_read = 0.0028 diff --git a/providers/novita-ai/models/xiaomimimo/mimo-v2.6-pro.toml b/providers/novita-ai/models/xiaomimimo/mimo-v2.6-pro.toml new file mode 100644 index 00000000000..13d76c65f11 --- /dev/null +++ b/providers/novita-ai/models/xiaomimimo/mimo-v2.6-pro.toml @@ -0,0 +1,13 @@ +# https://novita.ai/models/model-detail/xiaomimimo-mimo-v2.6-pro +# Toggle: thinking.type = enabled|disabled +base_model = "xiaomi/mimo-v2.6-pro" +structured_output = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.435 +output = 0.87 +cache_read = 0.0036 diff --git a/providers/novita-ai/models/zai-org/autoglm-phone-9b-multilingual.toml b/providers/novita-ai/models/zai-org/autoglm-phone-9b-multilingual.toml index 8a8347e2a94..4f519dec1b0 100644 --- a/providers/novita-ai/models/zai-org/autoglm-phone-9b-multilingual.toml +++ b/providers/novita-ai/models/zai-org/autoglm-phone-9b-multilingual.toml @@ -1,21 +1,6 @@ -name = "AutoGLM-Phone-9B-Multilingual" -description = "GLM vision model for visual reasoning, documents, and multimodal agents" -release_date = "2025-12-10" -last_updated = "2025-12-10" -attachment = true -reasoning = false -temperature = true -tool_call = false -open_weights = true +# https://novita.ai/models/model-detail/zai-org-autoglm-phone-9b-multilingual +base_model = "zhipuai/autoglm-phone-9b-multilingual" [cost] input = 0.035 output = 0.138 - -[limit] -context = 65_536 -output = 65_536 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-4.5-air.toml b/providers/novita-ai/models/zai-org/glm-4.5-air.toml index ec8f2f1d61b..a4cc2b5f5e2 100644 --- a/providers/novita-ai/models/zai-org/glm-4.5-air.toml +++ b/providers/novita-ai/models/zai-org/glm-4.5-air.toml @@ -1,25 +1,12 @@ -name = "GLM 4.5 Air" -description = "Efficient GLM model for fast reasoning, coding, and agent workflows" -family = "glm-air" -release_date = "2025-10-13" -last_updated = "2025-10-13" -attachment = false -reasoning = true +# https://novita.ai/models/model-detail/zai-org-glm-4.5-air +# Toggle: thinking.type = enabled|disabled +base_model = "zhipuai/glm-4.5-air" reasoning_options = [{ type = "toggle" }] -temperature = true -knowledge = "2025-04" -tool_call = true -open_weights = true + +[interleaved] +field = "reasoning_content" [cost] input = 0.13 output = 0.85 cache_read = 0.025 - -[limit] -context = 131_072 -output = 98_304 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-4.5.toml b/providers/novita-ai/models/zai-org/glm-4.5.toml deleted file mode 100644 index 55ddf8ee121..00000000000 --- a/providers/novita-ai/models/zai-org/glm-4.5.toml +++ /dev/null @@ -1,29 +0,0 @@ -name = "GLM-4.5" -description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering" -family = "glm" -release_date = "2025-07-28" -last_updated = "2025-07-28" -attachment = false -reasoning = true -temperature = true -tool_call = true -open_weights = true - -[[reasoning_options]] -type = "toggle" - -[cost] -input = 0.6 -output = 2.2 -cache_read = 0.11 - -[limit] -context = 131_072 -output = 98_304 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-4.5v.toml b/providers/novita-ai/models/zai-org/glm-4.5v.toml index 6a106f1cbd9..68e1b49445c 100644 --- a/providers/novita-ai/models/zai-org/glm-4.5v.toml +++ b/providers/novita-ai/models/zai-org/glm-4.5v.toml @@ -1,25 +1,16 @@ -name = "GLM 4.5V" -description = "GLM vision model for visual reasoning, documents, and multimodal agents" -family = "glmv" -release_date = "2025-08-11" -last_updated = "2025-08-11" -attachment = true -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -knowledge = "2025-04" -tool_call = true +# https://novita.ai/models/model-detail/zai-org-glm-4.5v +# Toggle: thinking.type = enabled|disabled +base_model = "zhipuai/glm-4.5v" structured_output = true -open_weights = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.6 output = 1.8 cache_read = 0.11 + [limit] context = 65_536 -output = 16_384 - -[modalities] -input = ["text", "video", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-4.6.toml b/providers/novita-ai/models/zai-org/glm-4.6.toml index 0e09c434022..cd254ba7a0e 100644 --- a/providers/novita-ai/models/zai-org/glm-4.6.toml +++ b/providers/novita-ai/models/zai-org/glm-4.6.toml @@ -1,28 +1,13 @@ -name = "GLM 4.6" -description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering" -family = "glm" -release_date = "2025-09-30" -last_updated = "2025-09-30" -attachment = false -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/zai-org-glm-4.6 +# Toggle: thinking.type = enabled|disabled +base_model = "zhipuai/glm-4.6" structured_output = true -open_weights = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.55 output = 2.2 cache_read = 0.11 - -[limit] -context = 204_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-4.6v.toml b/providers/novita-ai/models/zai-org/glm-4.6v.toml index f906825058d..de2b7b11bd1 100644 --- a/providers/novita-ai/models/zai-org/glm-4.6v.toml +++ b/providers/novita-ai/models/zai-org/glm-4.6v.toml @@ -1,16 +1,11 @@ -name = "GLM 4.6V" -description = "GLM vision model for visual reasoning, documents, and multimodal agents" -family = "glmv" -release_date = "2025-12-08" -last_updated = "2025-12-08" -attachment = true -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -knowledge = "2025-04" -tool_call = true +# https://novita.ai/models/model-detail/zai-org-glm-4.6v +# Toggle: thinking.type = enabled|disabled +base_model = "zhipuai/glm-4.6v" structured_output = true -open_weights = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.3 @@ -19,8 +14,3 @@ cache_read = 0.055 [limit] context = 131_072 -output = 32_768 - -[modalities] -input = ["text", "video", "image"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-4.7-flash.toml b/providers/novita-ai/models/zai-org/glm-4.7-flash.toml index c0079c3decf..482123b5a40 100644 --- a/providers/novita-ai/models/zai-org/glm-4.7-flash.toml +++ b/providers/novita-ai/models/zai-org/glm-4.7-flash.toml @@ -1,16 +1,11 @@ -name = "GLM-4.7-Flash" -description = "Efficient GLM model for fast reasoning, coding, and agent workflows" -family = "glm-flash" -release_date = "2026-01-19" -last_updated = "2026-01-19" -attachment = false -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/zai-org-glm-4.7-flash +# Toggle: thinking.type = enabled|disabled +base_model = "zhipuai/glm-4.7-flash" structured_output = true -knowledge = "2025-04" -open_weights = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.07 @@ -18,10 +13,4 @@ output = 0.4 cache_read = 0.01 [limit] -context = 200_000 output = 128_000 - - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-4.7.toml b/providers/novita-ai/models/zai-org/glm-4.7.toml index b8fc51214fd..fc503f63922 100644 --- a/providers/novita-ai/models/zai-org/glm-4.7.toml +++ b/providers/novita-ai/models/zai-org/glm-4.7.toml @@ -1,28 +1,13 @@ -name = "GLM-4.7" -description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering" -family = "glm" -release_date = "2025-12-22" -last_updated = "2025-12-22" -attachment = false -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/zai-org-glm-4.7 +# Toggle: thinking.type = enabled|disabled +base_model = "zhipuai/glm-4.7" structured_output = true -open_weights = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" [cost] input = 0.6 output = 2.2 cache_read = 0.11 - -[limit] -context = 204_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-5.1.toml b/providers/novita-ai/models/zai-org/glm-5.1.toml index 50d4249122e..2f00f4b5341 100644 --- a/providers/novita-ai/models/zai-org/glm-5.1.toml +++ b/providers/novita-ai/models/zai-org/glm-5.1.toml @@ -1,15 +1,10 @@ -name = "GLM-5.1" -description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering" -family = "glm" -release_date = "2026-03-27" -last_updated = "2026-03-27" -attachment = false -reasoning = true +# https://novita.ai/models/model-detail/zai-org-glm-5.1 +# Toggle: thinking.type = enabled|disabled +base_model = "zhipuai/glm-5.1" reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true -structured_output = true -open_weights = true + +[interleaved] +field = "reasoning_content" [cost] input = 1.38 @@ -18,11 +13,3 @@ cache_read = 0.26 [limit] context = 204_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/models/zai-org/glm-5.2.toml b/providers/novita-ai/models/zai-org/glm-5.2.toml index daa973cd32b..2b63ea088c1 100644 --- a/providers/novita-ai/models/zai-org/glm-5.2.toml +++ b/providers/novita-ai/models/zai-org/glm-5.2.toml @@ -1,4 +1,5 @@ -name = "GLM-5.2" +# https://novita.ai/models/model-detail/zai-org-glm-5.2 +# Effort: reasoning_effort = none|high|max; none suppresses reasoning. Native graded baseline retained. base_model = "zhipuai/glm-5.2" [[reasoning_options]] @@ -12,7 +13,6 @@ cache_read = 0.26 [limit] context = 1_048_576 -output = 131_072 [interleaved] field = "reasoning_content" diff --git a/providers/novita-ai/models/zai-org/glm-5.3-flash.toml b/providers/novita-ai/models/zai-org/glm-5.3-flash.toml new file mode 100644 index 00000000000..3b43c052e88 --- /dev/null +++ b/providers/novita-ai/models/zai-org/glm-5.3-flash.toml @@ -0,0 +1,18 @@ +# https://novita.ai/models/model-detail/zai-org-glm-5.3-flash +# Effort: reasoning_effort = low|high|max; thinking.type=disabled does not suppress reasoning on this route. +base_model = "zhipuai/glm-5.3-flash" +reasoning_options = [{ type = "effort", values = ["low", "high", "max"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 0.15 +output = 0.5 +cache_read = 0.03 + +[limit] +context = 1_048_576 + +[modalities] +input = ["text", "image", "video"] diff --git a/providers/novita-ai/models/zai-org/glm-5.3.toml b/providers/novita-ai/models/zai-org/glm-5.3.toml new file mode 100644 index 00000000000..b2dc7f8069a --- /dev/null +++ b/providers/novita-ai/models/zai-org/glm-5.3.toml @@ -0,0 +1,15 @@ +# https://novita.ai/models/model-detail/zai-org-glm-5.3 +# Effort: reasoning_effort = none|low|high|max; none suppresses reasoning. +base_model = "zhipuai/glm-5.3" +reasoning_options = [{ type = "effort", values = ["none", "low", "high", "max"] }] + +[interleaved] +field = "reasoning_content" + +[cost] +input = 1.4 +output = 4.4 +cache_read = 0.26 + +[limit] +context = 1_048_576 diff --git a/providers/novita-ai/models/zai-org/glm-5.toml b/providers/novita-ai/models/zai-org/glm-5.toml index 7b9123aa63a..bb681bc3d3a 100644 --- a/providers/novita-ai/models/zai-org/glm-5.toml +++ b/providers/novita-ai/models/zai-org/glm-5.toml @@ -1,28 +1,16 @@ -name = "GLM-5" -description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering" -family = "glm" -release_date = "2026-02-11" -last_updated = "2026-02-12" -attachment = false -reasoning = true -reasoning_options = [{ type = "toggle" }] -temperature = true -tool_call = true +# https://novita.ai/models/model-detail/zai-org-glm-5 +# Toggle: thinking.type = enabled|disabled +base_model = "zhipuai/glm-5" structured_output = true -open_weights = true +reasoning_options = [{ type = "toggle" }] + +[interleaved] +field = "reasoning_content" [cost] -input = 1.0 +input = 1 output = 3.2 cache_read = 0.2 [limit] context = 202_800 -output = 131_072 - -[interleaved] -field = "reasoning_content" - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/novita-ai/provider.toml b/providers/novita-ai/provider.toml index b7ebc291683..8d49dc3f5da 100644 --- a/providers/novita-ai/provider.toml +++ b/providers/novita-ai/provider.toml @@ -1,11 +1,9 @@ -name = "NovitaAI" +# Public catalog: https://novita.ai/models +# Inventory: https://api.novita.ai/openai/v1/models (account-scoped; absence is not retirement evidence). +# Chat reasoning controls are model-specific; see each route's leading wire comments. +# https://novita.ai/docs/api-reference/model-apis-llm-create-chat-completion +name = "Novita AI" env = ["NOVITA_API_KEY"] npm = "@ai-sdk/openai-compatible" -# Raw HTTP reasoning controls (sources accessed 2026-06-25): -# POST `/openai/v1/chat/completions` accepts top-level `enable_thinking = -# true|false` (default true), but documents it only for zai-org/glm-4.5 and -# deepseek/deepseek-v3.1, -v3.1-terminus, and -v3.2-exp. `separate_reasoning` -# is a distinct boolean documented only for deepseek/deepseek-r1-turbo. -# https://novita.ai/docs/api-reference/model-apis-llm-create-chat-completion doc = "https://novita.ai/docs/guides/introduction" api = "https://api.novita.ai/openai" From 91f2a57ffe568f60f2d0e455bc50d9f0224d58b8 Mon Sep 17 00:00:00 2001 From: Aiden Cline Date: Fri, 9 Oct 2026 11:47:08 -0500 Subject: [PATCH 2/4] fix(novita-ai): address catalog review findings --- providers/novita-ai/models/baai/bge-m3.toml | 4 +++- providers/novita-ai/models/baai/bge-reranker-v2-m3.toml | 3 ++- providers/novita-ai/models/baichuan/baichuan-m2-32b.toml | 6 ++++-- .../novita-ai/models/deepseek/deepseek-v4-flash-0731.toml | 4 ++-- providers/novita-ai/models/deepseek/deepseek-v4-flash.toml | 4 ++-- .../novita-ai/models/deepseek/deepseek-v4-pro-0813.toml | 4 ++-- providers/novita-ai/models/deepseek/deepseek-v4-pro.toml | 4 ++-- providers/novita-ai/models/mindai/macaron-v1-tall.toml | 6 +++--- .../novita-ai/models/minimax/minimax-m2.7-highspeed.toml | 1 - .../novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml | 6 +++--- providers/novita-ai/models/qwen/qwen3-reranker-8b.toml | 3 ++- 11 files changed, 25 insertions(+), 20 deletions(-) diff --git a/providers/novita-ai/models/baai/bge-m3.toml b/providers/novita-ai/models/baai/bge-m3.toml index 91b3d1736a6..18f0d20e873 100644 --- a/providers/novita-ai/models/baai/bge-m3.toml +++ b/providers/novita-ai/models/baai/bge-m3.toml @@ -1,7 +1,9 @@ # https://novita.ai/models/model-detail/baai-bge-m3 # POST /openai/v1/embeddings returns a 1024-dimensional vector; output is not a generative token limit. +# https://novita.ai/docs/api-reference/model-apis-llm-create-embeddings +# Input-only embedding workload: live usage reports completion_tokens=0. The generic catalog output-price field does not represent generated embedding tokens. base_model = "baai/bge-m3" [cost] input = 0.01 -output = 0.01 +output = 0 diff --git a/providers/novita-ai/models/baai/bge-reranker-v2-m3.toml b/providers/novita-ai/models/baai/bge-reranker-v2-m3.toml index 253c73be9b8..dd2a04fd758 100644 --- a/providers/novita-ai/models/baai/bge-reranker-v2-m3.toml +++ b/providers/novita-ai/models/baai/bge-reranker-v2-m3.toml @@ -1,10 +1,11 @@ # https://novita.ai/models/model-detail/baai-bge-reranker-v2-m3 # https://novita.ai/docs/api-reference/model-apis-llm-create-rerank +# Reranking returns relevance scores, not generated completion tokens; pricing is represented as an input-token rate. base_model = "baai/bge-reranker-v2-m3" [cost] input = 0.01 -output = 0.01 +output = 0 [limit] context = 8_000 diff --git a/providers/novita-ai/models/baichuan/baichuan-m2-32b.toml b/providers/novita-ai/models/baichuan/baichuan-m2-32b.toml index d81d7c37d03..76727e9f6e5 100644 --- a/providers/novita-ai/models/baichuan/baichuan-m2-32b.toml +++ b/providers/novita-ai/models/baichuan/baichuan-m2-32b.toml @@ -1,7 +1,9 @@ # https://novita.ai/models/model-detail/baichuan-baichuan-m2-32b -# Novita does not advertise reasoning for this route; preserve its non-reasoning serving mode, not the lab's broader thinking capability. +# https://huggingface.co/baichuan-inc/Baichuan-M2-32B +# Toggle: chat_template_kwargs.thinking_mode = on|off (native OpenAI-compatible serving baseline). +# Novita probes returned server_overload for default/on/off requests; the toggle is not live-verified on this host. Missing inventory features are not evidence that the checkpoint cannot reason. base_model = "baichuan/baichuan-m2-32b" -reasoning = false +reasoning_options = [{ type = "toggle" }] [cost] input = 0.07 diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-flash-0731.toml b/providers/novita-ai/models/deepseek/deepseek-v4-flash-0731.toml index 6bcc65c1fb8..67de5a4bd63 100644 --- a/providers/novita-ai/models/deepseek/deepseek-v4-flash-0731.toml +++ b/providers/novita-ai/models/deepseek/deepseek-v4-flash-0731.toml @@ -1,8 +1,8 @@ # https://novita.ai/models/model-detail/deepseek-deepseek-v4-flash-0731 # Toggle: thinking.type = enabled|disabled -# Effort: reasoning_effort = high|max (native baseline; low accepted as an alias). +# Effort: reasoning_effort = low|high|max (current same-surface baseline; all three values verified on Novita). base_model = "deepseek/deepseek-v4-flash-0731" -reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["high", "max"] }] +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "high", "max"] }] [interleaved] field = "reasoning_content" diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml b/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml index 687891efaf9..678dd10ca4d 100644 --- a/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml +++ b/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml @@ -1,9 +1,9 @@ # https://novita.ai/models/model-detail/deepseek-deepseek-v4-flash # Toggle: thinking.type = enabled|disabled -# Effort: reasoning_effort = high|max (native baseline; low accepted as an alias). +# Effort: reasoning_effort = low|high|max (current first-party baseline; low/max verified on Novita). # Explicit effort=high currently returns MODEL_NOT_FOUND, while default, low, and max requests work; retain the route. base_model = "deepseek/deepseek-v4-flash" -reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["high", "max"] }] +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "high", "max"] }] [interleaved] field = "reasoning_content" diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-pro-0813.toml b/providers/novita-ai/models/deepseek/deepseek-v4-pro-0813.toml index b0f7129b90a..d407b562edf 100644 --- a/providers/novita-ai/models/deepseek/deepseek-v4-pro-0813.toml +++ b/providers/novita-ai/models/deepseek/deepseek-v4-pro-0813.toml @@ -1,8 +1,8 @@ # https://novita.ai/models/model-detail/deepseek-deepseek-v4-pro-0813 # Toggle: thinking.type = enabled|disabled -# Effort: reasoning_effort = high|max (native baseline; low accepted as an alias). +# Effort: reasoning_effort = low|high|max (current first-party baseline; all three values verified on Novita). base_model = "deepseek/deepseek-v4-pro-0813" -reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["high", "max"] }] +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "high", "max"] }] [interleaved] field = "reasoning_content" diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-pro.toml b/providers/novita-ai/models/deepseek/deepseek-v4-pro.toml index 1b0c165e35e..b4666115b00 100644 --- a/providers/novita-ai/models/deepseek/deepseek-v4-pro.toml +++ b/providers/novita-ai/models/deepseek/deepseek-v4-pro.toml @@ -1,8 +1,8 @@ # https://novita.ai/models/model-detail/deepseek-deepseek-v4-pro # Toggle: thinking.type = enabled|disabled -# Effort: reasoning_effort = high|max (native baseline; low accepted as an alias). +# Effort: reasoning_effort = low|high|max (current first-party baseline; all three values verified on Novita). base_model = "deepseek/deepseek-v4-pro" -reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["high", "max"] }] +reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "high", "max"] }] [interleaved] field = "reasoning_content" diff --git a/providers/novita-ai/models/mindai/macaron-v1-tall.toml b/providers/novita-ai/models/mindai/macaron-v1-tall.toml index 390c9f5da46..daf13be40c3 100644 --- a/providers/novita-ai/models/mindai/macaron-v1-tall.toml +++ b/providers/novita-ai/models/mindai/macaron-v1-tall.toml @@ -1,9 +1,9 @@ # https://novita.ai/models/model-detail/mindai-macaron-v1-tall -# Toggle: reasoning_effort = none|high (off|on); this field, not thinking.type, suppresses reasoning. -# Other accepted effort values were not demonstrated to be distinct depth controls, so only the verified binary control is represented. +# Effort: reasoning_effort = none|high; this field, not thinking.type, suppresses reasoning with none. +# Other accepted effort values were not demonstrated to be distinct depth controls, so only the verified off/on effort values are represented. # thinking.type=disabled does not suppress reasoning on this route; thinking_budget is not honored. base_model = "mindlab/macaron-v1-tall" -reasoning_options = [{ type = "toggle" }] +reasoning_options = [{ type = "effort", values = ["none", "high"] }] [interleaved] field = "reasoning_content" diff --git a/providers/novita-ai/models/minimax/minimax-m2.7-highspeed.toml b/providers/novita-ai/models/minimax/minimax-m2.7-highspeed.toml index d23af2e2cfa..bfa348f0767 100644 --- a/providers/novita-ai/models/minimax/minimax-m2.7-highspeed.toml +++ b/providers/novita-ai/models/minimax/minimax-m2.7-highspeed.toml @@ -3,7 +3,6 @@ # Preserve curated pricing: /models does not return this route. base_model = "minimax/MiniMax-M2.7-highspeed" reasoning_options = [] -last_updated = "2026-05-27" structured_output = true [interleaved] diff --git a/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml b/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml index b2c88e0adf9..2d15766cdb6 100644 --- a/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml +++ b/providers/novita-ai/models/qwen/qwen3-omni-30b-a3b-thinking.toml @@ -1,9 +1,9 @@ # https://novita.ai/models/model-detail/qwen-qwen3-omni-30b-a3b-thinking -# Live requests emit no reasoning by default or when disabled; thinking.type=enabled and enable_thinking=true are rejected. -# This route currently does not expose the underlying checkpoint's thinking mode. +# https://huggingface.co/Qwen/Qwen3-Omni-30B-A3B-Thinking +# Native thinking-only checkpoint; no caller toggle/budget is established. Novita rejects the tested enable/effort controls and emits no separate reasoning on the tested prompts, which does not establish that the checkpoint lacks reasoning capability. # Preserve the curated input-audio rate, which the inventory does not return; output-audio pricing is inapplicable to this text-output checkpoint. base_model = "alibaba/qwen3-omni-30b-a3b-thinking" -reasoning = false +reasoning_options = [] structured_output = true [cost] diff --git a/providers/novita-ai/models/qwen/qwen3-reranker-8b.toml b/providers/novita-ai/models/qwen/qwen3-reranker-8b.toml index 355f43757fd..0eedf51b452 100644 --- a/providers/novita-ai/models/qwen/qwen3-reranker-8b.toml +++ b/providers/novita-ai/models/qwen/qwen3-reranker-8b.toml @@ -1,8 +1,9 @@ # https://novita.ai/models/model-detail/qwen-qwen3-reranker-8b # https://novita.ai/docs/api-reference/model-apis-llm-create-rerank # Public description identifies the text Qwen3 Reranker 8B; the site's HF mirror URL points to a different VL checkpoint. +# Reranking returns relevance scores, not generated completion tokens; pricing is represented as an input-token rate. base_model = "alibaba/qwen3-reranker-8b" [cost] input = 0.05 -output = 0.05 +output = 0 From 8231414029185503ec49dcae6a8f8a410e1e75f7 Mon Sep 17 00:00:00 2001 From: Aiden Cline Date: Fri, 9 Oct 2026 12:12:02 -0500 Subject: [PATCH 3/4] fix(novita-ai): remove retired flash route and cite output limit --- .../models/deepseek/deepseek-v4-flash.toml | 18 ------------------ .../novita-ai/models/minimax/minimax-m2.5.toml | 1 + 2 files changed, 1 insertion(+), 18 deletions(-) delete mode 100644 providers/novita-ai/models/deepseek/deepseek-v4-flash.toml diff --git a/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml b/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml deleted file mode 100644 index 678dd10ca4d..00000000000 --- a/providers/novita-ai/models/deepseek/deepseek-v4-flash.toml +++ /dev/null @@ -1,18 +0,0 @@ -# https://novita.ai/models/model-detail/deepseek-deepseek-v4-flash -# Toggle: thinking.type = enabled|disabled -# Effort: reasoning_effort = low|high|max (current first-party baseline; low/max verified on Novita). -# Explicit effort=high currently returns MODEL_NOT_FOUND, while default, low, and max requests work; retain the route. -base_model = "deepseek/deepseek-v4-flash" -reasoning_options = [{ type = "toggle" }, { type = "effort", values = ["low", "high", "max"] }] - -[interleaved] -field = "reasoning_content" - -[cost] -input = 0.14 -output = 0.28 -cache_read = 0.028 - -[limit] -context = 1_048_576 -output = 393_216 diff --git a/providers/novita-ai/models/minimax/minimax-m2.5.toml b/providers/novita-ai/models/minimax/minimax-m2.5.toml index c4b6b440c72..4cc204f75b4 100644 --- a/providers/novita-ai/models/minimax/minimax-m2.5.toml +++ b/providers/novita-ai/models/minimax/minimax-m2.5.toml @@ -1,4 +1,5 @@ # https://novita.ai/models/model-detail/minimax-minimax-m2.5 +# Novita's /models reports max_output_tokens=131100, and the linked model page's example uses max_tokens=131100 (verified 2026-10-09); this is a host delta, not a transcription of the lab's 131072 limit. # Always-on reasoning: thinking.type, enable_thinking, chat_template_kwargs.enable_thinking, and effort=none did not suppress it. base_model = "minimax/MiniMax-M2.5" structured_output = true From aa2ca040ed9b3cf6bf11aa2e9f9850ae779a39a8 Mon Sep 17 00:00:00 2001 From: Aiden Cline Date: Fri, 9 Oct 2026 12:18:32 -0500 Subject: [PATCH 4/4] fix(novita-ai): defer case-sensitive Sao10K offerings --- models/sao10k/l3-70b-euryale-v2.1.toml | 24 ------------------- models/sao10k/l3-8b-stheno-v3.2.toml | 24 ------------------- .../models/Sao10K/L3-8B-Stheno-v3.2.toml | 10 -------- .../models/sao10k/l3-70b-euryale-v2.1.toml | 8 ------- 4 files changed, 66 deletions(-) delete mode 100644 models/sao10k/l3-70b-euryale-v2.1.toml delete mode 100644 models/sao10k/l3-8b-stheno-v3.2.toml delete mode 100644 providers/novita-ai/models/Sao10K/L3-8B-Stheno-v3.2.toml delete mode 100644 providers/novita-ai/models/sao10k/l3-70b-euryale-v2.1.toml diff --git a/models/sao10k/l3-70b-euryale-v2.1.toml b/models/sao10k/l3-70b-euryale-v2.1.toml deleted file mode 100644 index aec4c3084fc..00000000000 --- a/models/sao10k/l3-70b-euryale-v2.1.toml +++ /dev/null @@ -1,24 +0,0 @@ -# https://huggingface.co/Sao10K/L3-70B-Euryale-v2.1 -name = "L3 70B Euryale V2.1" -description = "Llama 3 70B fine-tune for creative dialogue and roleplay" -family = "llama" -release_date = "2024-06-11" -last_updated = "2024-06-11" -attachment = false -reasoning = false -tool_call = false -temperature = true -open_weights = true -license = "CC-BY-NC-4.0" - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] - -[[weights]] -label = "Hugging Face" -url = "https://huggingface.co/Sao10K/L3-70B-Euryale-v2.1" diff --git a/models/sao10k/l3-8b-stheno-v3.2.toml b/models/sao10k/l3-8b-stheno-v3.2.toml deleted file mode 100644 index 3fa3cc5bbf4..00000000000 --- a/models/sao10k/l3-8b-stheno-v3.2.toml +++ /dev/null @@ -1,24 +0,0 @@ -# https://huggingface.co/Sao10K/L3-8B-Stheno-v3.2 -name = "L3 8B Stheno V3.2" -description = "Llama 3 8B fine-tune for creative writing and roleplay" -family = "llama" -release_date = "2024-06-05" -last_updated = "2024-06-05" -attachment = false -reasoning = false -tool_call = false -temperature = true -open_weights = true -license = "CC-BY-NC-4.0" - -[limit] -context = 8_192 -output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] - -[[weights]] -label = "Hugging Face" -url = "https://huggingface.co/Sao10K/L3-8B-Stheno-v3.2" diff --git a/providers/novita-ai/models/Sao10K/L3-8B-Stheno-v3.2.toml b/providers/novita-ai/models/Sao10K/L3-8B-Stheno-v3.2.toml deleted file mode 100644 index ba9d3b4c938..00000000000 --- a/providers/novita-ai/models/Sao10K/L3-8B-Stheno-v3.2.toml +++ /dev/null @@ -1,10 +0,0 @@ -# https://novita.ai/models/model-detail/Sao10K-L3-8B-Stheno-v3.2 -# Preserve the exact case-sensitive model ID returned by Novita. -base_model = "sao10k/l3-8b-stheno-v3.2" - -[cost] -input = 0.05 -output = 0.05 - -[limit] -output = 32_000 diff --git a/providers/novita-ai/models/sao10k/l3-70b-euryale-v2.1.toml b/providers/novita-ai/models/sao10k/l3-70b-euryale-v2.1.toml deleted file mode 100644 index 899f57fb037..00000000000 --- a/providers/novita-ai/models/sao10k/l3-70b-euryale-v2.1.toml +++ /dev/null @@ -1,8 +0,0 @@ -# https://novita.ai/models/model-detail/sao10k-l3-70b-euryale-v2.1 -# Preserve the exact case-sensitive model ID returned by Novita. -base_model = "sao10k/l3-70b-euryale-v2.1" -tool_call = true - -[cost] -input = 1.48 -output = 1.48