From 9e5bed6b4462eb0a8eb6df6981de007f8c38d265 Mon Sep 17 00:00:00 2001
From: "opencode-agent[bot]" <opencode-agent[bot]@users.noreply.github.com>
Date: Fri, 9 Oct 2026 18:32:51 +0000
Subject: [PATCH] chore(sync): update Kilo model catalog

---
 .../models/deepseek/deepseek-r1-0528.toml     | 19 +------------
 .../models/deepseek/deepseek-v4-flash.toml    |  6 ++---
 .../meta-llama/llama-3.2-1b-instruct.toml     | 13 +--------
 .../kilo/models/moonshotai/kimi-k2-0905.toml  | 18 +------------
 .../models/moonshotai/kimi-k2-thinking.toml   |  3 ++-
 .../kilo/models/moonshotai/kimi-k2.6.toml     |  4 +--
 .../models/qwen/qwen-plus-2025-07-28.toml     | 23 ----------------
 providers/kilo/models/qwen/qwen3-14b.toml     |  4 +--
 .../models/qwen/qwen3-235b-a22b-2507.toml     |  4 +--
 .../qwen/qwen3-235b-a22b-thinking-2507.toml   | 10 +++----
 .../kilo/models/qwen/qwen3-235b-a22b.toml     | 14 ----------
 .../qwen/qwen3-30b-a3b-instruct-2507.toml     |  4 +--
 .../qwen/qwen3-30b-a3b-thinking-2507.toml     | 27 -------------------
 providers/kilo/models/qwen/qwen3-30b-a3b.toml |  4 +--
 providers/kilo/models/qwen/qwen3-8b.toml      | 27 -------------------
 .../qwen/qwen3-coder-30b-a3b-instruct.toml    |  4 +--
 .../kilo/models/qwen/qwen3-coder-next.toml    |  5 ++--
 .../kilo/models/qwen/qwen3-coder-plus.toml    | 12 ---------
 providers/kilo/models/qwen/qwen3-coder.toml   |  5 ++--
 .../kilo/models/qwen/qwen3-max-thinking.toml  | 27 -------------------
 providers/kilo/models/qwen/qwen3-max.toml     |  9 -------
 .../qwen/qwen3-next-80b-a3b-instruct.toml     |  5 ++--
 .../qwen/qwen3-next-80b-a3b-thinking.toml     |  4 +++
 .../qwen/qwen3-vl-235b-a22b-instruct.toml     |  5 ++--
 .../qwen/qwen3-vl-235b-a22b-thinking.toml     | 10 -------
 .../qwen/qwen3-vl-30b-a3b-instruct.toml       | 16 +++--------
 .../qwen/qwen3-vl-30b-a3b-thinking.toml       | 10 +++----
 .../models/qwen/qwen3-vl-32b-instruct.toml    | 23 ----------------
 .../models/qwen/qwen3-vl-8b-instruct.toml     | 11 ++++----
 .../models/qwen/qwen3-vl-8b-thinking.toml     | 27 -------------------
 .../kilo/models/qwen/qwen3.6-max-preview.toml | 12 ---------
 providers/kilo/models/tencent/hy3.toml        |  6 ++---
 .../kilo/models/tencent/hy4-preview.toml      |  6 ++---
 .../~deepseek/deepseek-flash-latest.toml      |  4 +--
 .../~deepseek/deepseek-v4-flash-latest.toml   |  6 ++---
 .../kilo/models/~moonshotai/kimi-latest.toml  |  6 ++---
 .../kilo/models/~z-ai/glm-flash-latest.toml   |  2 +-
 37 files changed, 70 insertions(+), 325 deletions(-)
 delete mode 100644 providers/kilo/models/qwen/qwen-plus-2025-07-28.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-235b-a22b.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-30b-a3b-thinking-2507.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-8b.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-coder-plus.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-max-thinking.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-max.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-vl-235b-a22b-thinking.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-vl-32b-instruct.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3-vl-8b-thinking.toml
 delete mode 100644 providers/kilo/models/qwen/qwen3.6-max-preview.toml

diff --git a/providers/kilo/models/deepseek/deepseek-r1-0528.toml b/providers/kilo/models/deepseek/deepseek-r1-0528.toml
index 6f834792a3a..55c29817339 100644
--- a/providers/kilo/models/deepseek/deepseek-r1-0528.toml
+++ b/providers/kilo/models/deepseek/deepseek-r1-0528.toml
@@ -1,14 +1,5 @@
-name = "DeepSeek: R1 0528"
+base_model = "deepseek/deepseek-r1-0528"
 description = "DeepSeek reasoning model for multi-step analysis, math, coding, and tools"
-family = "deepseek"
-release_date = "2025-05-28"
-last_updated = "2025-05-28"
-attachment = false
-reasoning = true
-temperature = true
-tool_call = true
-structured_output = true
-open_weights = false
 
 [[reasoning_options]]
 type = "effort"
@@ -18,11 +9,3 @@ values = ["high"]
 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/kilo/models/deepseek/deepseek-v4-flash.toml b/providers/kilo/models/deepseek/deepseek-v4-flash.toml
index 1f32751fb70..406b2e8b785 100644
--- a/providers/kilo/models/deepseek/deepseek-v4-flash.toml
+++ b/providers/kilo/models/deepseek/deepseek-v4-flash.toml
@@ -6,9 +6,9 @@ type = "effort"
 values = ["none", "high", "xhigh"]
 
 [cost]
-input = 0.14
-output = 0.28
-cache_read = 0.028
+input = 0.0131
+output = 1.28
+cache_read = 0.0131
 
 [limit]
 context = 1_048_576
diff --git a/providers/kilo/models/meta-llama/llama-3.2-1b-instruct.toml b/providers/kilo/models/meta-llama/llama-3.2-1b-instruct.toml
index 6c7a0c72120..cce4b71c294 100644
--- a/providers/kilo/models/meta-llama/llama-3.2-1b-instruct.toml
+++ b/providers/kilo/models/meta-llama/llama-3.2-1b-instruct.toml
@@ -1,14 +1,7 @@
-name = "Meta: Llama 3.2 1B Instruct"
+base_model = "meta/llama-3.2-1b-instruct"
 description = "Open Llama instruction model for multilingual chat, reasoning, and coding"
-family = "llama"
-release_date = "2024-09-25"
-last_updated = "2024-09-25"
-attachment = false
-reasoning = false
-temperature = true
 tool_call = false
 structured_output = false
-open_weights = false
 
 [cost]
 input = 0.027
@@ -17,7 +10,3 @@ output = 0.201
 [limit]
 context = 60_000
 output = 54_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/kilo/models/moonshotai/kimi-k2-0905.toml b/providers/kilo/models/moonshotai/kimi-k2-0905.toml
index 39cad5b17ea..adb34cfb895 100644
--- a/providers/kilo/models/moonshotai/kimi-k2-0905.toml
+++ b/providers/kilo/models/moonshotai/kimi-k2-0905.toml
@@ -1,23 +1,7 @@
-name = "MoonshotAI: Kimi K2 0905"
+base_model = "moonshotai/kimi-k2-0905"
 description = "Kimi model for long-context chat, coding, and agentic reasoning"
-family = "kimi-k2"
-release_date = "2025-09-04"
-last_updated = "2025-09-04"
-attachment = false
-reasoning = false
-temperature = true
-tool_call = true
 structured_output = true
-open_weights = false
 
 [cost]
 input = 0.6
 output = 2.5
-
-[limit]
-context = 262_144
-output = 98_304
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/kilo/models/moonshotai/kimi-k2-thinking.toml b/providers/kilo/models/moonshotai/kimi-k2-thinking.toml
index f90682d224c..c0ea113422d 100644
--- a/providers/kilo/models/moonshotai/kimi-k2-thinking.toml
+++ b/providers/kilo/models/moonshotai/kimi-k2-thinking.toml
@@ -9,6 +9,7 @@ values = ["high"]
 [cost]
 input = 0.6
 output = 2.5
+cache_read = 0.15
 
 [limit]
-output = 235_929
+output = 98_304
diff --git a/providers/kilo/models/moonshotai/kimi-k2.6.toml b/providers/kilo/models/moonshotai/kimi-k2.6.toml
index 2407c57290b..85924561b3f 100644
--- a/providers/kilo/models/moonshotai/kimi-k2.6.toml
+++ b/providers/kilo/models/moonshotai/kimi-k2.6.toml
@@ -6,9 +6,9 @@ type = "effort"
 values = ["none", "high"]
 
 [cost]
-input = 0.4348
+input = 0.4344
 output = 2.45
-cache_read = 0.1202
+cache_read = 0.1077
 
 [limit]
 output = 235_929
diff --git a/providers/kilo/models/qwen/qwen-plus-2025-07-28.toml b/providers/kilo/models/qwen/qwen-plus-2025-07-28.toml
deleted file mode 100644
index eae43d98aad..00000000000
--- a/providers/kilo/models/qwen/qwen-plus-2025-07-28.toml
+++ /dev/null
@@ -1,23 +0,0 @@
-name = "Qwen: Qwen Plus 0728 (retires Oct 9)"
-description = "Qwen instruction model for multilingual chat, reasoning, and tool use"
-family = "qwen"
-release_date = "2025-09-08"
-last_updated = "2025-09-08"
-attachment = false
-reasoning = false
-temperature = true
-tool_call = true
-structured_output = true
-open_weights = false
-
-[cost]
-input = 0.26
-output = 0.78
-
-[limit]
-context = 1_000_000
-output = 32_768
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/kilo/models/qwen/qwen3-14b.toml b/providers/kilo/models/qwen/qwen3-14b.toml
index 2ce8bc46042..de1cf69ac7a 100644
--- a/providers/kilo/models/qwen/qwen3-14b.toml
+++ b/providers/kilo/models/qwen/qwen3-14b.toml
@@ -15,8 +15,8 @@ type = "effort"
 values = ["none", "high"]
 
 [cost]
-input = 0.2275
-output = 0.91
+input = 0.12
+output = 0.24
 
 [limit]
 context = 40_960
diff --git a/providers/kilo/models/qwen/qwen3-235b-a22b-2507.toml b/providers/kilo/models/qwen/qwen3-235b-a22b-2507.toml
index 3f37ded3629..7ff01ac8b33 100644
--- a/providers/kilo/models/qwen/qwen3-235b-a22b-2507.toml
+++ b/providers/kilo/models/qwen/qwen3-235b-a22b-2507.toml
@@ -11,8 +11,8 @@ structured_output = true
 open_weights = false
 
 [cost]
-input = 0.1495
-output = 0.598
+input = 0.09
+output = 0.55
 
 [limit]
 context = 262_144
diff --git a/providers/kilo/models/qwen/qwen3-235b-a22b-thinking-2507.toml b/providers/kilo/models/qwen/qwen3-235b-a22b-thinking-2507.toml
index e0603b155cc..036a1287d4a 100644
--- a/providers/kilo/models/qwen/qwen3-235b-a22b-thinking-2507.toml
+++ b/providers/kilo/models/qwen/qwen3-235b-a22b-thinking-2507.toml
@@ -1,4 +1,4 @@
-name = "Qwen: Qwen3 235B A22B Thinking 2507 (retires Oct 9)"
+name = "Qwen: Qwen3 235B A22B Thinking 2507"
 description = "Qwen reasoning model for deliberate problem solving, math, and coding"
 family = "qwen"
 release_date = "2025-07-25"
@@ -15,12 +15,12 @@ type = "effort"
 values = ["high"]
 
 [cost]
-input = 0.23
-output = 2.3
+input = 0.45
+output = 3.5
 
 [limit]
-context = 131_072
-output = 117_964
+context = 128_000
+output = 16_384
 
 [modalities]
 input = ["text"]
diff --git a/providers/kilo/models/qwen/qwen3-235b-a22b.toml b/providers/kilo/models/qwen/qwen3-235b-a22b.toml
deleted file mode 100644
index 66286de6054..00000000000
--- a/providers/kilo/models/qwen/qwen3-235b-a22b.toml
+++ /dev/null
@@ -1,14 +0,0 @@
-base_model = "alibaba/qwen3-235b-a22b"
-description = "Qwen instruction model for multilingual chat, reasoning, and tool use"
-structured_output = false
-
-[[reasoning_options]]
-type = "effort"
-values = ["none", "high"]
-
-[cost]
-input = 0.455
-output = 1.82
-
-[limit]
-output = 8_192
diff --git a/providers/kilo/models/qwen/qwen3-30b-a3b-instruct-2507.toml b/providers/kilo/models/qwen/qwen3-30b-a3b-instruct-2507.toml
index 1da6f655c94..c655d188f55 100644
--- a/providers/kilo/models/qwen/qwen3-30b-a3b-instruct-2507.toml
+++ b/providers/kilo/models/qwen/qwen3-30b-a3b-instruct-2507.toml
@@ -11,8 +11,8 @@ structured_output = true
 open_weights = false
 
 [cost]
-input = 0.13
-output = 0.52
+input = 0.1
+output = 0.3
 
 [limit]
 context = 262_144
diff --git a/providers/kilo/models/qwen/qwen3-30b-a3b-thinking-2507.toml b/providers/kilo/models/qwen/qwen3-30b-a3b-thinking-2507.toml
deleted file mode 100644
index 97b8a302fc7..00000000000
--- a/providers/kilo/models/qwen/qwen3-30b-a3b-thinking-2507.toml
+++ /dev/null
@@ -1,27 +0,0 @@
-name = "Qwen: Qwen3 30B A3B Thinking 2507 (retires Oct 9)"
-description = "Qwen reasoning model for deliberate problem solving, math, and coding"
-family = "qwen"
-release_date = "2025-08-28"
-last_updated = "2025-08-28"
-attachment = false
-reasoning = true
-temperature = true
-tool_call = true
-structured_output = false
-open_weights = false
-
-[[reasoning_options]]
-type = "effort"
-values = ["high"]
-
-[cost]
-input = 0.2
-output = 2.4
-
-[limit]
-context = 81_920
-output = 32_768
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/kilo/models/qwen/qwen3-30b-a3b.toml b/providers/kilo/models/qwen/qwen3-30b-a3b.toml
index a8c6cef56d6..d4852076bc9 100644
--- a/providers/kilo/models/qwen/qwen3-30b-a3b.toml
+++ b/providers/kilo/models/qwen/qwen3-30b-a3b.toml
@@ -7,8 +7,8 @@ type = "effort"
 values = ["none", "high"]
 
 [cost]
-input = 0.13
-output = 0.52
+input = 0.12
+output = 0.5
 
 [limit]
 context = 40_960
diff --git a/providers/kilo/models/qwen/qwen3-8b.toml b/providers/kilo/models/qwen/qwen3-8b.toml
deleted file mode 100644
index d6560afcb2b..00000000000
--- a/providers/kilo/models/qwen/qwen3-8b.toml
+++ /dev/null
@@ -1,27 +0,0 @@
-name = "Qwen: Qwen3 8B (retires Oct 9)"
-description = "Qwen instruction model for multilingual chat, reasoning, and tool use"
-family = "qwen"
-release_date = "2025-04-28"
-last_updated = "2025-04-28"
-attachment = false
-reasoning = true
-temperature = true
-tool_call = true
-structured_output = false
-open_weights = false
-
-[[reasoning_options]]
-type = "effort"
-values = ["none", "high"]
-
-[cost]
-input = 0.117
-output = 0.455
-
-[limit]
-context = 131_072
-output = 8_192
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/kilo/models/qwen/qwen3-coder-30b-a3b-instruct.toml b/providers/kilo/models/qwen/qwen3-coder-30b-a3b-instruct.toml
index c227fd2e1bb..1eea4094a1e 100644
--- a/providers/kilo/models/qwen/qwen3-coder-30b-a3b-instruct.toml
+++ b/providers/kilo/models/qwen/qwen3-coder-30b-a3b-instruct.toml
@@ -3,8 +3,8 @@ description = "Qwen coding model for software agents, repository edits, and code
 structured_output = true
 
 [cost]
-input = 0.2925
-output = 1.4625
+input = 0.07
+output = 0.28
 
 [limit]
 output = 235_929
diff --git a/providers/kilo/models/qwen/qwen3-coder-next.toml b/providers/kilo/models/qwen/qwen3-coder-next.toml
index 55ce6e1ccd7..679e6f18d9b 100644
--- a/providers/kilo/models/qwen/qwen3-coder-next.toml
+++ b/providers/kilo/models/qwen/qwen3-coder-next.toml
@@ -2,8 +2,9 @@ base_model = "alibaba/qwen3-coder-next"
 description = "Qwen coding model for software agents, repository edits, and code reasoning"
 
 [cost]
-input = 0.3
-output = 1.5
+input = 0.12
+output = 0.8
+cache_read = 0.07
 
 [limit]
 output = 235_929
diff --git a/providers/kilo/models/qwen/qwen3-coder-plus.toml b/providers/kilo/models/qwen/qwen3-coder-plus.toml
deleted file mode 100644
index 77d3c155cec..00000000000
--- a/providers/kilo/models/qwen/qwen3-coder-plus.toml
+++ /dev/null
@@ -1,12 +0,0 @@
-base_model = "alibaba/qwen3-coder-plus"
-description = "Qwen coding model for software agents, repository edits, and code reasoning"
-structured_output = true
-
-[cost]
-input = 0.65
-output = 3.25
-cache_read = 0.13
-cache_write = 0.8125
-
-[limit]
-context = 1_000_000
diff --git a/providers/kilo/models/qwen/qwen3-coder.toml b/providers/kilo/models/qwen/qwen3-coder.toml
index 1719e16ce21..b3950b5ff6d 100644
--- a/providers/kilo/models/qwen/qwen3-coder.toml
+++ b/providers/kilo/models/qwen/qwen3-coder.toml
@@ -11,8 +11,9 @@ structured_output = true
 open_weights = false
 
 [cost]
-input = 0.975
-output = 4.875
+input = 0.3
+output = 1
+cache_read = 0.1
 
 [limit]
 context = 262_144
diff --git a/providers/kilo/models/qwen/qwen3-max-thinking.toml b/providers/kilo/models/qwen/qwen3-max-thinking.toml
deleted file mode 100644
index 3d9d2977a64..00000000000
--- a/providers/kilo/models/qwen/qwen3-max-thinking.toml
+++ /dev/null
@@ -1,27 +0,0 @@
-name = "Qwen: Qwen3 Max Thinking (retires Oct 9)"
-description = "Qwen reasoning model for deliberate problem solving, math, and coding"
-family = "qwen"
-release_date = "2026-02-09"
-last_updated = "2026-02-09"
-attachment = false
-reasoning = true
-temperature = true
-tool_call = true
-structured_output = true
-open_weights = false
-
-[[reasoning_options]]
-type = "effort"
-values = ["none", "high"]
-
-[cost]
-input = 0.78
-output = 3.9
-
-[limit]
-context = 262_144
-output = 65_536
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/kilo/models/qwen/qwen3-max.toml b/providers/kilo/models/qwen/qwen3-max.toml
deleted file mode 100644
index 59a284e4a04..00000000000
--- a/providers/kilo/models/qwen/qwen3-max.toml
+++ /dev/null
@@ -1,9 +0,0 @@
-base_model = "alibaba/qwen3-max"
-description = "Flagship Qwen model for complex reasoning, coding, and agentic workflows"
-structured_output = true
-
-[cost]
-input = 0.78
-output = 3.9
-cache_read = 0.156
-cache_write = 0.975
diff --git a/providers/kilo/models/qwen/qwen3-next-80b-a3b-instruct.toml b/providers/kilo/models/qwen/qwen3-next-80b-a3b-instruct.toml
index 8247cd4ca01..0505204d268 100644
--- a/providers/kilo/models/qwen/qwen3-next-80b-a3b-instruct.toml
+++ b/providers/kilo/models/qwen/qwen3-next-80b-a3b-instruct.toml
@@ -3,8 +3,9 @@ description = "Qwen instruction model for multilingual chat, reasoning, and tool
 structured_output = true
 
 [cost]
-input = 0.0975
-output = 0.78
+input = 0.1
+output = 1.1
+cache_read = 0.07
 
 [limit]
 context = 262_144
diff --git a/providers/kilo/models/qwen/qwen3-next-80b-a3b-thinking.toml b/providers/kilo/models/qwen/qwen3-next-80b-a3b-thinking.toml
index 7c76fb082af..5d67b671b7a 100644
--- a/providers/kilo/models/qwen/qwen3-next-80b-a3b-thinking.toml
+++ b/providers/kilo/models/qwen/qwen3-next-80b-a3b-thinking.toml
@@ -9,3 +9,7 @@ values = ["high"]
 [cost]
 input = 0.15
 output = 1.2
+
+[limit]
+context = 262_144
+output = 235_929
diff --git a/providers/kilo/models/qwen/qwen3-vl-235b-a22b-instruct.toml b/providers/kilo/models/qwen/qwen3-vl-235b-a22b-instruct.toml
index 6ba5414d7f2..957d575f2c1 100644
--- a/providers/kilo/models/qwen/qwen3-vl-235b-a22b-instruct.toml
+++ b/providers/kilo/models/qwen/qwen3-vl-235b-a22b-instruct.toml
@@ -2,5 +2,6 @@ base_model = "alibaba/qwen3-vl-235b-a22b-instruct"
 description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
 
 [cost]
-input = 0.26
-output = 1.04
+input = 0.21
+output = 1.9
+cache_read = 0.1
diff --git a/providers/kilo/models/qwen/qwen3-vl-235b-a22b-thinking.toml b/providers/kilo/models/qwen/qwen3-vl-235b-a22b-thinking.toml
deleted file mode 100644
index 029c19f51c3..00000000000
--- a/providers/kilo/models/qwen/qwen3-vl-235b-a22b-thinking.toml
+++ /dev/null
@@ -1,10 +0,0 @@
-base_model = "alibaba/qwen3-vl-235b-a22b-thinking"
-description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
-
-[[reasoning_options]]
-type = "effort"
-values = ["high"]
-
-[cost]
-input = 0.4
-output = 4
diff --git a/providers/kilo/models/qwen/qwen3-vl-30b-a3b-instruct.toml b/providers/kilo/models/qwen/qwen3-vl-30b-a3b-instruct.toml
index 83d4e34b47c..2117029a6e4 100644
--- a/providers/kilo/models/qwen/qwen3-vl-30b-a3b-instruct.toml
+++ b/providers/kilo/models/qwen/qwen3-vl-30b-a3b-instruct.toml
@@ -1,23 +1,13 @@
-name = "Qwen: Qwen3 VL 30B A3B Instruct"
+base_model = "alibaba/qwen3-vl-30b-a3b-instruct"
 description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
-family = "qwen"
-release_date = "2025-10-06"
-last_updated = "2025-10-06"
-attachment = true
-reasoning = false
-temperature = true
-tool_call = true
 structured_output = true
-open_weights = false
 
 [cost]
-input = 0.13
-output = 0.52
+input = 0.15
+output = 0.6
 
 [limit]
-context = 262_144
 output = 16_384
 
 [modalities]
 input = ["text", "image"]
-output = ["text"]
diff --git a/providers/kilo/models/qwen/qwen3-vl-30b-a3b-thinking.toml b/providers/kilo/models/qwen/qwen3-vl-30b-a3b-thinking.toml
index 5634fd3caf4..136589d1757 100644
--- a/providers/kilo/models/qwen/qwen3-vl-30b-a3b-thinking.toml
+++ b/providers/kilo/models/qwen/qwen3-vl-30b-a3b-thinking.toml
@@ -1,4 +1,4 @@
-name = "Qwen: Qwen3 VL 30B A3B Thinking (retires Oct 9)"
+name = "Qwen: Qwen3 VL 30B A3B Thinking"
 description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
 family = "qwen"
 release_date = "2025-10-06"
@@ -15,12 +15,12 @@ type = "effort"
 values = ["high"]
 
 [cost]
-input = 0.2
-output = 2.4
+input = 0.29
+output = 1
 
 [limit]
-context = 131_072
-output = 32_768
+context = 262_144
+output = 235_929
 
 [modalities]
 input = ["text", "image"]
diff --git a/providers/kilo/models/qwen/qwen3-vl-32b-instruct.toml b/providers/kilo/models/qwen/qwen3-vl-32b-instruct.toml
deleted file mode 100644
index 71b89f4f7e6..00000000000
--- a/providers/kilo/models/qwen/qwen3-vl-32b-instruct.toml
+++ /dev/null
@@ -1,23 +0,0 @@
-name = "Qwen: Qwen3 VL 32B Instruct (retires Oct 9)"
-description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
-family = "qwen"
-release_date = "2025-10-23"
-last_updated = "2025-10-23"
-attachment = true
-reasoning = false
-temperature = true
-tool_call = true
-structured_output = true
-open_weights = false
-
-[cost]
-input = 0.104
-output = 0.416
-
-[limit]
-context = 131_072
-output = 32_768
-
-[modalities]
-input = ["text", "image"]
-output = ["text"]
diff --git a/providers/kilo/models/qwen/qwen3-vl-8b-instruct.toml b/providers/kilo/models/qwen/qwen3-vl-8b-instruct.toml
index 1526026df06..a2bae1a5925 100644
--- a/providers/kilo/models/qwen/qwen3-vl-8b-instruct.toml
+++ b/providers/kilo/models/qwen/qwen3-vl-8b-instruct.toml
@@ -1,4 +1,4 @@
-name = "Qwen: Qwen3 VL 8B Instruct (retires Oct 9)"
+name = "Qwen: Qwen3 VL 8B Instruct"
 description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
 family = "qwen"
 release_date = "2025-10-14"
@@ -11,12 +11,13 @@ structured_output = true
 open_weights = false
 
 [cost]
-input = 0.117
-output = 0.455
+input = 0.25
+output = 0.75
+cache_read = 0.12
 
 [limit]
-context = 131_072
-output = 32_768
+context = 262_144
+output = 235_929
 
 [modalities]
 input = ["image", "text"]
diff --git a/providers/kilo/models/qwen/qwen3-vl-8b-thinking.toml b/providers/kilo/models/qwen/qwen3-vl-8b-thinking.toml
deleted file mode 100644
index 7874c34fb09..00000000000
--- a/providers/kilo/models/qwen/qwen3-vl-8b-thinking.toml
+++ /dev/null
@@ -1,27 +0,0 @@
-name = "Qwen: Qwen3 VL 8B Thinking (retires Oct 9)"
-description = "Qwen vision-language model for visual reasoning, documents, and agent tasks"
-family = "qwen"
-release_date = "2025-10-14"
-last_updated = "2025-10-14"
-attachment = true
-reasoning = true
-temperature = true
-tool_call = true
-structured_output = true
-open_weights = false
-
-[[reasoning_options]]
-type = "effort"
-values = ["high"]
-
-[cost]
-input = 0.18
-output = 2.1
-
-[limit]
-context = 131_072
-output = 32_768
-
-[modalities]
-input = ["image", "text"]
-output = ["text"]
diff --git a/providers/kilo/models/qwen/qwen3.6-max-preview.toml b/providers/kilo/models/qwen/qwen3.6-max-preview.toml
deleted file mode 100644
index c45be57de87..00000000000
--- a/providers/kilo/models/qwen/qwen3.6-max-preview.toml
+++ /dev/null
@@ -1,12 +0,0 @@
-base_model = "alibaba/qwen3.6-max-preview"
-description = "Flagship Qwen model for complex reasoning, coding, and agentic workflows"
-structured_output = true
-
-[[reasoning_options]]
-type = "effort"
-values = ["none", "high"]
-
-[cost]
-input = 1.027
-output = 6.162
-cache_write = 1.28375
diff --git a/providers/kilo/models/tencent/hy3.toml b/providers/kilo/models/tencent/hy3.toml
index 2bdfa3cbd7b..e96e3110ab1 100644
--- a/providers/kilo/models/tencent/hy3.toml
+++ b/providers/kilo/models/tencent/hy3.toml
@@ -7,9 +7,9 @@ type = "effort"
 values = ["none", "low", "high"]
 
 [cost]
-input = 0.132
-output = 0.528
-cache_read = 0.033
+input = 0.0825
+output = 0.33
+cache_read = 0.020625
 
 [limit]
 context = 262_144
diff --git a/providers/kilo/models/tencent/hy4-preview.toml b/providers/kilo/models/tencent/hy4-preview.toml
index 664c533e3b8..fc75aa842da 100644
--- a/providers/kilo/models/tencent/hy4-preview.toml
+++ b/providers/kilo/models/tencent/hy4-preview.toml
@@ -7,9 +7,9 @@ type = "effort"
 values = ["none", "low", "high"]
 
 [cost]
-input = 0.834
-output = 2.501
-cache_read = 0.042
+input = 0.7506
+output = 2.2509
+cache_read = 0.0378
 
 [limit]
 context = 1_048_576
diff --git a/providers/kilo/models/~deepseek/deepseek-flash-latest.toml b/providers/kilo/models/~deepseek/deepseek-flash-latest.toml
index d1c3f01de04..9c65c2a2f72 100644
--- a/providers/kilo/models/~deepseek/deepseek-flash-latest.toml
+++ b/providers/kilo/models/~deepseek/deepseek-flash-latest.toml
@@ -16,8 +16,8 @@ values = ["none", "low", "high", "max"]
 
 [cost]
 input = 0.016
-output = 1.2
-cache_read = 0.01
+output = 0.6
+cache_read = 0.005
 
 [limit]
 context = 1_048_576
diff --git a/providers/kilo/models/~deepseek/deepseek-v4-flash-latest.toml b/providers/kilo/models/~deepseek/deepseek-v4-flash-latest.toml
index 609aa877da7..5ffdf5532e9 100644
--- a/providers/kilo/models/~deepseek/deepseek-v4-flash-latest.toml
+++ b/providers/kilo/models/~deepseek/deepseek-v4-flash-latest.toml
@@ -15,9 +15,9 @@ type = "effort"
 values = ["none", "low", "high", "max"]
 
 [cost]
-input = 0.0079
-output = 1.28
-cache_read = 0.0079
+input = 0.00512
+output = 0.01467
+cache_read = 0.00512
 
 [limit]
 context = 1_048_576
diff --git a/providers/kilo/models/~moonshotai/kimi-latest.toml b/providers/kilo/models/~moonshotai/kimi-latest.toml
index ca9a3475392..034d6a4f66c 100644
--- a/providers/kilo/models/~moonshotai/kimi-latest.toml
+++ b/providers/kilo/models/~moonshotai/kimi-latest.toml
@@ -15,9 +15,9 @@ type = "effort"
 values = ["none", "low", "high", "max"]
 
 [cost]
-input = 0.49
-output = 13
-cache_read = 0.3
+input = 0.45
+output = 14
+cache_read = 0.29
 
 [limit]
 context = 1_048_576
diff --git a/providers/kilo/models/~z-ai/glm-flash-latest.toml b/providers/kilo/models/~z-ai/glm-flash-latest.toml
index e44252aa6a6..5da90f40923 100644
--- a/providers/kilo/models/~z-ai/glm-flash-latest.toml
+++ b/providers/kilo/models/~z-ai/glm-flash-latest.toml
@@ -16,7 +16,7 @@ values = ["low", "high", "max"]
 
 [cost]
 input = 0.032
-output = 0.264718
+output = 0.140682
 cache_read = 0.02
 
 [limit]