List your company

Rationale Support Quality and Alignment on MedQA

95.65Accuracy (Q)GRPO
72.530878.532984.53590.5371Jun 12, 2026

Evaluation results

MethodLinks
GRPOModel=Ministral 3 8B2026.0695.655.79
CoRAModel=Ministral 3 8B2026.0694.995.08
BaseModel=Ministral 3 8B2026.0694.96.09
SFTModel=Ministral 3 8B2026.0694.885.35
CoRAModel=Gemma 2 9B2026.0694.476.02
GRPOModel=Gemma 2 9B2026.0682.6316.91
SFTModel=Gemma 2 9B2026.0682.5917.14
BaseModel=Gemma 2 9B2026.0681.4718.31
CoRAModel=Qwen2.5 7B2026.0679.8720.29
SFTModel=Qwen2.5 7B2026.0677.6721.76
GRPOModel=Qwen2.5 7B2026.0674.7722.31
BaseModel=Qwen2.5 7B2026.0673.4226.2