List your company
ResearchFormerly WizWand

Medical Reasoning on MMLU Pro

80.7AccuracyMDAgents
21.31636.73352.1567.567Aug 11, 2025Oct 3, 2025Nov 25, 2025Jan 17, 2026Mar 11, 2026May 3, 2026Jun 25, 2026
Updated 4mo ago

Evaluation results

MethodLinks
MDAgentsBase Model=GPT-4o2025.0880.7
DyLANBase Model=GPT-4o2025.0880
TMA-AllComponBase Model=GPT-4o2025.0879.7
ReConcileBase Model=GPT-4o2025.0878.8
m1kModel=Qwen2.5-7B-Instruct, Training=Finetuning, Sample size=1000 examples2026.0664.5
Middle PerplexityModel=Qwen2.5-7B-Instruct, Training=Finetuning, Sample size=1000 examples2026.0663.6
OursModel=Qwen2.5-7B-Instruct, Training=Finetuning, Sample size=1000 examples2026.0663
RandomModel=Qwen2.5-7B-Instruct, Training=Finetuning, Sample size=1000 examples2026.0662.7
Embedding DiversityModel=Qwen2.5-7B-Instruct, Training=Finetuning, Sample size=1000 examples2026.0662.5
S2LModel=Qwen2.5-7B-Instruct, Training=Finetuning, Sample size=1000 examples2026.0662.5
LearnabilityModel=Qwen2.5-7B-Instruct, Training=Finetuning, Sample size=1000 examples2026.0660.8
MedAgentsBase Model=GPT-4o2025.0846.7
TMA-AllComponBase Model=MedGemma-4B2025.0846.4
ReConcileBase Model=Gemma-3-4B2025.0835.8
TMA-AllComponBase Model=Gemma-3-4B2025.0835.6
DyLANBase Model=Gemma-3-4B2025.0827.4
MDAgentsBase Model=Gemma-3-4B2025.0824.4
MedAgentsBase Model=Gemma-3-4B2025.0823.6