From 506a634284531a7f15f22dc3bfeb76d078c53a0b Mon Sep 17 00:00:00 2001 From: Claude Date: Thu, 25 Jun 2026 01:17:24 +0000 Subject: [PATCH] Default batch_eval workflow to groq/gpt-oss-120b for Pass 2 The workflow_dispatch dropdown and EVAL_MODEL fallback still pointed at the deprecated llama-3.3-70b-versatile, so manual GitHub Actions runs kept hitting the free-tier limits config.py's default already avoids. --- .github/workflows/batch_eval.yml | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/.github/workflows/batch_eval.yml b/.github/workflows/batch_eval.yml index 22035b7..2fe9078 100644 --- a/.github/workflows/batch_eval.yml +++ b/.github/workflows/batch_eval.yml @@ -11,6 +11,7 @@ on: description: "Pass 2 identification model" type: choice options: + - groq/gpt-oss-120b - llama-3.3-70b-versatile - cerebras/gpt-oss-120b - cerebras/zai-glm-4.7 @@ -19,7 +20,7 @@ on: - gemini-2.0-flash - mistral-large-latest - mistral-medium-3 - default: "llama-3.3-70b-versatile" + default: "groq/gpt-oss-120b" required: false judge_model: description: "Pass 4 judge model (cross-family recommended)" @@ -29,6 +30,7 @@ on: - gemini-2.5-flash - gemini-2.0-flash - gemini-2.5-pro + - groq/gpt-oss-120b - llama-3.3-70b-versatile - cerebras/gpt-oss-120b - mistral-large-latest @@ -81,7 +83,7 @@ jobs: GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }} CEREBRAS_API_KEY: ${{ secrets.CEREBRAS_API_KEY }} MISTRAL_API_KEY: ${{ secrets.MISTRAL_API_KEY }} - EVAL_MODEL: ${{ inputs.eval_model || 'llama-3.3-70b-versatile' }} + EVAL_MODEL: ${{ inputs.eval_model || 'groq/gpt-oss-120b' }} JUDGE_MODEL: ${{ inputs.judge_model == 'auto' && '' || inputs.judge_model }} TRIAGE_MODEL: ${{ inputs.triage_model || 'cerebras/gpt-oss-120b' }}