List your company

Medical Visual Question Answering (Multiple-Choice) on IU-XRay (test)

96.05AccuracyLIAR
-3.84222.091548.02573.9585May 11, 2026

Evaluation results

MethodLinks
LIARModel=GPT-5, Retriever=BGE-VL2026.0596.05
Clean RAG (avg.)Model=GPT-5, Retriever=Average2026.0594.2
M³AttModel=GPT-5, Retriever=BGE-VL2026.0591.73
LIARModel=GPT-4o, Retriever=CLIP2026.0589.89
Clean RAG (avg.)Model=GPT-4o, Retriever=Average2026.0589.63
LIARModel=GPT-4o, Retriever=BGE-VL2026.0588.6
w/o RAGModel=GPT-4o, Retriever=None2026.0587.41
LIARModel=GPT-5, Retriever=CLIP2026.0587.21
M³AttModel=GPT-4o, Retriever=CLIP2026.0586.52
M³AttModel=GPT-4o, Retriever=BGE-VL2026.0585.93
w/o RAGModel=GPT-5, Retriever=None2026.0584.44
M³AttModel=GPT-5, Retriever=CLIP2026.0584.29
LIARModel=GPT-4o, Retriever=Filtered2026.0581.53
LIARModel=GPT-5, Retriever=Filtered2026.0580.76
M³AttModel=GPT-4o, Retriever=Filtered2026.0577.04
M³AttModel=GPT-5, Retriever=Filtered2026.0576.29
LIARModel=Claude-4.5, Retriever=CLIP2026.0573
LIARModel=Claude-4.5, Retriever=BGE-VL2026.0568.92
Clean RAG (avg.)Model=Claude-4.5, Retriever=Average2026.0567.69
M³AttModel=Claude-4.5, Retriever=CLIP2026.0567.6
Clean RAG (avg.)Model=Gemini-2.5, Retriever=Average2026.0566.28
w/o RAGModel=Claude-4.5, Retriever=None2026.0565.92
LIARModel=Gemini-2.5, Retriever=CLIP2026.0565.17
M³AttModel=Claude-4.5, Retriever=BGE-VL2026.0565.07
LIARModel=Gemini-2.5, Retriever=BGE-VL2026.0560.65
M³AttModel=Gemini-2.5, Retriever=CLIP2026.0559.47
M³AttModel=Gemini-2.5, Retriever=BGE-VL2026.0556.52
LIARModel=Claude-4.5, Retriever=Filtered2026.0555.44
M³AttModel=Claude-4.5, Retriever=Filtered2026.0551.11
w/o RAGModel=Gemini-2.5, Retriever=None2026.0549.63
LIARModel=LLaVA-Med, Retriever=CLIP2026.0548.04
LIARModel=Gemini-2.5, Retriever=Filtered2026.0544.83
M³AttModel=Gemini-2.5, Retriever=Filtered2026.0541.48
M³AttModel=LLaVA-Med, Retriever=CLIP2026.0536.96
Clean RAG (avg.)Model=LLaVA-Med, Retriever=Average2026.0525.77
LIARModel=LLaVA-Med, Retriever=BGE-VL2026.054.51
M³AttModel=LLaVA-Med, Retriever=BGE-VL2026.053.51
LIARModel=LLaVA-Med, Retriever=Filtered2026.051.56
w/o RAGModel=LLaVA-Med, Retriever=None2026.050
M³AttModel=LLaVA-Med, Retriever=Filtered2026.050