[LLM] models update
This commit is contained in:
@@ -180,6 +180,26 @@
|
||||
"maxTokens": 2048,
|
||||
"cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }
|
||||
},
|
||||
{
|
||||
"id": "GLM-4.7-Flash-UD-Q3_K_XL",
|
||||
"name": "GLM-4.7 Flash Q3 · 24k — fast coder (~28 t/s)",
|
||||
"reasoning": true,
|
||||
"compat": { "thinkingFormat": "qwen-chat-template" },
|
||||
"input": ["text"],
|
||||
"contextWindow": 24576,
|
||||
"maxTokens": 2048,
|
||||
"cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }
|
||||
},
|
||||
{
|
||||
"id": "Ornith-1.5-35B-A3B-IQ3_XXS",
|
||||
"name": "Ornith 1.5 35B · 24k · vision — agentic coder, thinking (~32 t/s)",
|
||||
"reasoning": true,
|
||||
"compat": { "thinkingFormat": "qwen-chat-template" },
|
||||
"input": ["text", "image"],
|
||||
"contextWindow": 24576,
|
||||
"maxTokens": 2048,
|
||||
"cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }
|
||||
},
|
||||
{
|
||||
"id": "gpt-oss-20b",
|
||||
"name": "gpt-oss 20B · 64k — fast reasoning + tools (~38 t/s)",
|
||||
@@ -295,6 +315,26 @@
|
||||
"maxTokens": 2048,
|
||||
"cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }
|
||||
},
|
||||
{
|
||||
"id": "GLM-4.7-Flash-UD-Q3_K_XL",
|
||||
"name": "GLM-4.7 Flash Q3 · 24k — fast coder (~28 t/s)",
|
||||
"reasoning": true,
|
||||
"compat": { "thinkingFormat": "qwen-chat-template" },
|
||||
"input": ["text"],
|
||||
"contextWindow": 24576,
|
||||
"maxTokens": 2048,
|
||||
"cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }
|
||||
},
|
||||
{
|
||||
"id": "Ornith-1.5-35B-A3B-IQ3_XXS",
|
||||
"name": "Ornith 1.5 35B · 24k · vision — agentic coder, thinking (~32 t/s)",
|
||||
"reasoning": true,
|
||||
"compat": { "thinkingFormat": "qwen-chat-template" },
|
||||
"input": ["text", "image"],
|
||||
"contextWindow": 24576,
|
||||
"maxTokens": 2048,
|
||||
"cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 }
|
||||
},
|
||||
{
|
||||
"id": "gpt-oss-20b",
|
||||
"name": "gpt-oss 20B · 64k — fast reasoning + tools (~38 t/s)",
|
||||
|
||||
Reference in New Issue
Block a user