List your company
ResearchFormerly WizWand

Multi-modal Question Answering on MedXpertQA-MM

62.4AccuracyQwen3.5-27B
4.274419.364734.45549.5453Aug 4, 2025Sep 14, 2025Oct 25, 2025Dec 6, 2025Jan 16, 2026Feb 26, 2026Apr 9, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Qwen3.5-27BMode=REASONING, Architecture=Dense, # Total Params=27B, # Activated Params=27B2026.0462.4
o1Model Category=Proprietary2026.0149.7
Qwen3-VL-235B-A22BMode=Thinking, Architecture=MoE, # Total Params=236B, # Activated Params=23B2026.0447.6
EXAONE 4.5 33BMode=REASONING, Architecture=Dense, # Total Params=33B, # Activated Params=33B2026.0442.1
Qwen3-VL-32BMode=Thinking, Architecture=Dense, # Total Params=33B, # Activated Params=33B2026.0441.6
Gemini2.5-proModel Category=Proprietary2026.0139.5
PulseMind-72BModel Category=Open-source, Model Scale=~72B2026.0136.7
GPT-4oDecoding=Greedy2025.0835.95
MedVLThinker-32B RLDecoding=Greedy2025.0834.6
GPT-5 miniMode=REASONING: HIGH, Architecture=-, # Total Params=-, # Activated Params=-2026.0434.4
Lingshu-32BModel Category=Open-source, Model Scale=~32B2026.0130.9
Gemme 3 27BDecoding=Greedy2025.0830.8
PulseMind-32BModel Category=Open-source, Model Scale=~32B2026.0129.6
GPT-4o-miniDecoding=Greedy2025.0828.55
Qwen2.5-VL-32B-InstructDecoding=Greedy2025.0827.68
Qwen2.5VL-72BModel Category=Open-source, Model Scale=~72B2026.0127.6
InternVL3-78BModel Category=Open-source, Model Scale=~72B2026.0127.4
InternVL3-38BModel Category=Open-source, Model Scale=~32B2026.0125.2
Qwen2.5VL-32BModel Category=Open-source, Model Scale=~32B2026.0125.2
MedVLThinker-7B RLDecoding=Greedy2025.0824.43
SPINEBase Model=Qwen2.5-VL-3B-Instruct2025.1123.84
MedVLThinker-3B RLDecoding=Greedy2025.0822.9
TTRLBase Model=Qwen2.5-VL-3B-Instruct2025.1122.61
Llava Med v1.5 Mistral 7BDecoding=Greedy2025.0822.56
GPT-4oModel Category=Proprietary2026.0122.3
LMSIBase Model=Qwen2.5-VL-3B-Instruct2025.1122.01
HuatuoGPT-Vision-7BDecoding=Greedy2025.0822
Gemme 3 4BDecoding=Greedy2025.0821.89
HuatuoGPT-Vision-34BDecoding=Greedy2025.0821.8
Qwen2.5-VL-3B-InstructDecoding=Greedy2025.0820.69
Self-ConsistencyBase Model=Qwen2.5-VL-3B-Instruct2025.1119.47
Qwen2.5-VL-7B-InstructDecoding=Greedy2025.0818.89
HuatuoGPT-vision-34BModel Category=Open-source, Model Scale=~32B2026.0117.3
No adaptationBase Model=Qwen2.5-VL-3B-Instruct2025.1117.17
LLAVA-med-34BModel Category=Open-source, Model Scale=~32B2026.0116.4
MedGemma 27BDecoding=Greedy2025.0812.13
MedGemma 4BDecoding=Greedy2025.088.17
SEALONGBase Model=Qwen2.5-VL-3B-Instruct2025.116.51