diff --git a/packages/gateway/src/routes/completion.ts b/packages/gateway/src/routes/completion.ts index 3e4fb2b..f52a966 100644 --- a/packages/gateway/src/routes/completion.ts +++ b/packages/gateway/src/routes/completion.ts @@ -449,7 +449,7 @@ async function executeCompletion(body: CompletionRequest, startMs: number, callI if (action === 'llm_judge') { try { const verdict = await llmJudge(body.input, { - model: 'qwen2.5:3b', + model: process.env['LLM_JUDGE_MODEL'] || 'qwen2.5:3b', callLLM: async (req) => { const resp = await callOllama( { model: req.model, prompt: req.prompt, system: req.system, stream: false, options: { temperature: 0, num_predict: 8, ...(req.options ?? {}) } },