chore(sync): update Kilo model catalog

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"]

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

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"]

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"]

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

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

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

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

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"]

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

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

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

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

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

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

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

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