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

Medical Visual Question Answering on PMC-VQA

62Accuracy

GPT-5

Updated 5d ago

Evaluation Results

Method	Links
GPT-5 2026.01		62
Gemini-2.5-pro 2026.01		62
OpenAI-o3 2026.04		61.5
GPT-5 2026.04		60.2
GPT-5-mini 2026.01		60
Claude-4.5-haiku 2026.01		59
Claude-4.5-sonnet 2026.01		59
GEMINI-3-FLASH 2026.01		58.1
MEDVISTA-R1 2026.01		58
GPT-5 2026.01		57.7
LINGSHU-7B + MED-SCOUT 2026.01		57.4
InternVL3-8B 2026.01		56.5
MedAgents 2025.08		56.5
TMA-AllCompon 2025.08		56.4
MDAgents 2025.08		56.4
MedTutor-R1 2025.12		56.3
LINGSHU-7B 2026.01		56.3
MEDIC-AD 2026.03		56.1
GPT-o4-mini 2026.01		56
Cold-start w/o Reasoning 2026.01		56
Lingshu 2026.03		56
DyLAN 2025.08		56
ReConcile 2025.08		56
HUATUOGPT-VISION-7B + MED-SCOUT 2026.01		55.9
Gemini 2.5 Pro 2026.04		55.9
Citrus-V 2026.03		55.6
GPT-4.1 2026.04		55.2
MedLVR 2026.04		53.6
MedTutor-R1 w/ LLaVA-based 2025.12		53.09
HUATUOGPT-VISION-7B 2026.01		53
InternVL3-8B 2026.04		52.7
Qwen3vl-8B 2026.01		52.5
MedTutor-R1 w/o RL 2025.12		52.28
INTERNVL3-8B 2026.01		52
QWEN2.5-VL-7B-INSTRUCT 2026.01		51.8
GRPO w/o Tools 2026.01		51.5
HuatuoGPT-Vision-34B 2026.04		51.4
Cold-start w/o Tools 2026.01		51
Qwen2.5-VL-32B 2026.04		50.4
QWEN2.5-VL-3B-INSTRUCT 2026.01		50.2
MEDVISTAGYM (Qwen3vl-8B) 2026.01		49.5
MEDVISTA-R1 2026.01		49.5
MEDVISTAGYM 2026.04		49.5
Qwen2.5vl-7B 2026.01		49
Direct GRPO w/o cold-start 2026.01		49
Qwen2.5-VL-7B 2026.04		49
MEDGEMMA-4B-IT 2026.01		48.7
GRPO w/o Tools 2026.01		48.5
Qwen2.5VL 2025.12		48.15
Cold-start w/o Reasoning 2026.01		47.5
Direct GRPO w/o cold-start 2026.01		47
Cold-start w/o Tools 2026.01		46.5
Med-R1 2026.04		45.8
QWEN3-VL-8B-INSTRUCT + MED-SCOUT 2026.01		45.5
QWEN3-VL-4B-INSTRUCT + MED-SCOUT 2026.01		45.1
MedVLM-R1 2026.04		44.8
Internvl3-2B 2026.01		44.5
QWEN3-VL-8B-INSTRUCT 2026.01		43.9
MEDVISTAGYM (InternVL3-8B) 2026.01		43.5
TMA-AllCompon 2025.08		43.2
QWEN3-VL-4B-INSTRUCT 2026.01		42.8
MEDVISTAGYM (InternVL3-2B) 2026.01		40
TMA-AllCompon 2025.08		38.5
LLaVA-Next-13B 2026.04		36.6
LLaVA-v1.5-8B 2026.04		36.4
LLaVA-Next-7B 2026.04		35.5
Gemini-2.5-flash 2026.01		35
LLAVA-MED-7B 2026.01		32.4
LLaVA-Med 2026.03		30.5
LLaVA-Med-7B 2026.01		27
RadFM 2026.04		25.9
LLaVA-Med-7B 2026.04		24.7
Med-Flamingo 2026.04		23.3
Lingshu 2026.03		6.4