Disable gpt-oss.ai.unturf.com endpoint (returns 404)

This commit is contained in:
russell@unturf.com 2026-01-27 08:12:07 -05:00
parent 519e281623
commit ab13f2c19f

View file

@ -178,16 +178,7 @@ export const VLLM_ENDPOINTS = [
'hf.co/unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF:Q4_K_M': 32768,
}
},
{
id: "gpt-oss.ai.unturf.com",
url: "https://gpt-oss.ai.unturf.com/v1",
// Ollama endpoints don't return max_model_len, so we specify it here
// Note: Model supports 262144 but limited by GPU VRAM (24GB)
modelContextWindows: {
'gpt-oss:latest': 64000,
'deepseek-r1:14b-qwen-distill-q8_0': 64000,
}
},
// { id: "gpt-oss.ai.unturf.com", url: "https://gpt-oss.ai.unturf.com/v1" }, // Disabled - upstream returns 404
];
// Helper to get vault value or localStorage fallback