List your company

Multi-Turn Medical Dialogue on MedQA

68.69AccuracyMEDIQ
26.767637.651348.53559.4187Feb 10, 2026

Evaluation results

MethodLinks
MEDIQModel=Gemini-2.5-Pro, Method Type=Prompt2026.0268.69
ATPO (U1 + U2)Model=Qwen3-8B, Method Type=SFT+RL2026.0264.07
MEDIQModel=GPT-4o, Method Type=Prompt2026.0263.15
ATPO (U1)Model=Qwen3-8B, Method Type=SFT+RL2026.0262.57
TreePOModel=Qwen3-8B, Method Type=SFT+RL2026.0261.81
GRPOModel=Qwen3-8B, Method Type=SFT+RL2026.0257.92
PPO (H-MDP)Model=Qwen3-8B, Method Type=SFT+RL2026.0257.81
PPO (MDP)Model=Qwen3-8B, Method Type=SFT+RL2026.0257.38
ATPO (U1 + U2)Model=Qwen3-4B, Method Type=SFT+RL2026.0255.47
SFTModel=Qwen3-8B, Method Type=SFT2026.0253.75
TreePOModel=Qwen3-4B, Method Type=SFT+RL2026.0253.74
ATPO (U1)Model=Qwen3-4B, Method Type=SFT+RL2026.0253.15
GRPOModel=Qwen3-4B, Method Type=SFT+RL2026.0251.17
PPO (MDP)Model=Qwen3-4B, Method Type=SFT+RL2026.0250.6
DFTModel=Qwen3-8B, Method Type=SFT2026.0248.8
PPO (H-MDP)Model=Qwen3-4B, Method Type=SFT+RL2026.0248.58
SFTModel=Qwen3-4B, Method Type=SFT2026.0247.15
MEDIQModel=Qwen3-8B, Method Type=Prompt2026.0246.03
DirectModel=Qwen3-8B, Method Type=Prompt2026.0245.22
DirectModel=Qwen3-4B, Method Type=Prompt2026.0244.94
ATPO (U1 + U2)Model=Qwen3-1.7B, Method Type=SFT+RL2026.0242.87
ATPO (U1)Model=Qwen3-1.7B, Method Type=SFT+RL2026.0242.54
TreePOModel=Qwen3-1.7B, Method Type=SFT+RL2026.0242.05
DFTModel=Qwen3-4B, Method Type=SFT2026.0241.72
GRPOModel=Qwen3-1.7B, Method Type=SFT+RL2026.0241.17
MEDIQModel=Qwen3-4B, Method Type=Prompt2026.0240.11
PPO (H-MDP)Model=Qwen3-1.7B, Method Type=SFT+RL2026.0239.08
PPO (MDP)Model=Qwen3-1.7B, Method Type=SFT+RL2026.0238.64
MEDIQModel=Qwen3-1.7B, Method Type=Prompt2026.0234.2
DirectModel=Qwen3-1.7B, Method Type=Prompt2026.0234.05
SFTModel=Qwen3-1.7B, Method Type=SFT2026.0233.42
DFTModel=Qwen3-1.7B, Method Type=SFT2026.0228.38