Share your thoughts, 1 month free Claude Pro on usSee more

Multi-modal Question Answering on MedXpertQA-MM

62.4Accuracy

Qwen3.5-27B

Updated 6d ago

Evaluation Results

Method	Links
Qwen3.5-27B 2026.04		62.4
o1 2026.01		49.7
Qwen3-VL-235B-A22B 2026.04		47.6
EXAONE 4.5 33B 2026.04		42.1
Qwen3-VL-32B 2026.04		41.6
Gemini2.5-pro 2026.01		39.5
PulseMind-72B 2026.01		36.7
GPT-4o 2025.08		35.95
MedVLThinker-32B RL 2025.08		34.6
GPT-5 mini 2026.04		34.4
Lingshu-32B 2026.01		30.9
Gemme 3 27B 2025.08		30.8
PulseMind-32B 2026.01		29.6
GPT-4o-mini 2025.08		28.55
Qwen2.5-VL-32B-Instruct 2025.08		27.68
Qwen2.5VL-72B 2026.01		27.6
InternVL3-78B 2026.01		27.4
InternVL3-38B 2026.01		25.2
Qwen2.5VL-32B 2026.01		25.2
MedVLThinker-7B RL 2025.08		24.43
SPINE 2025.11		23.84
MedVLThinker-3B RL 2025.08		22.9
TTRL 2025.11		22.61
Llava Med v1.5 Mistral 7B 2025.08		22.56
GPT-4o 2026.01		22.3
LMSI 2025.11		22.01
HuatuoGPT-Vision-7B 2025.08		22
Gemme 3 4B 2025.08		21.89
HuatuoGPT-Vision-34B 2025.08		21.8
Qwen2.5-VL-3B-Instruct 2025.08		20.69
Self-Consistency 2025.11		19.47
Qwen2.5-VL-7B-Instruct 2025.08		18.89
HuatuoGPT-vision-34B 2026.01		17.3
No adaptation 2025.11		17.17
LLAVA-med-34B 2026.01		16.4
MedGemma 27B 2025.08		12.13
MedGemma 4B 2025.08		8.17
SEALONG 2025.11		6.51