Visual Question Answering on BreastStage-Bench Open-ended benchmark leaderboard · SOTA2 Research

Visual Question Answering on BreastStage-Bench Open-ended

95.38Screening (CT) ScoreQwen3-VL (SFT)

Evaluation Results

Method	Links
Qwen3-VL (SFT)2026.06		95.38	94.43	94.57	91.7	95.28	88.36	57.96	88.24
BreastGPT (cluster)2026.06		95.29	95.97	95.48	93.24	95.72	89.93	63.8	89.92
BreastGPT (learn)2026.06		94.73	95.86	87.12	93.57	85.22	81.78	63.36	85.95
MedDr2026.06		64.52	61.46	67.68	62.83	73.03	64.86	50.46	63.55
RadFM2026.06		61.07	50.51	67.63	50.8	69.15	66.76	50.44	59.48
Hulu-Med2026.06		60.84	58.98	75.46	55.95	72.28	59.48	50.18	61.88
InternVL3.52026.06		58.05	50.74	56.56	48.8	58.89	55.94	46.5	53.64
LLaVA-OneVision-1.52026.06		56.06	—	59.84	—	59.19	58.12	45.25	55.69
MiMo-VL2026.06		55.98	42.49	63.7	41.69	65.66	63.08	39.86	53.21
Qwen3-VL2026.06		55.49	45.2	47.71	44.93	51.46	48.98	41.26	47.86
HuatuoGPT-V2026.06		51.47	51.86	55.68	49.2	55.08	52.93	45.78	51.71
Gemini-3.1-Pro2026.06		51.46	51.19	41.6	48.94	41.39	42.8	45.77	46.16
HealthGPT2026.06		50.17	54.66	53.08	51.92	59.38	55.27	46.41	52.98
Lingshu2026.06		49.84	52.6	53.41	48.95	53.76	49.6	43.69	50.26
GPT-5.42026.06		49.34	53.43	59.85	50.42	59.2	58.07	44.73	53.58
Qwen3-VL2026.06		48.53	44.94	44.85	44.34	46.24	44.15	41.21	44.89
Qwen2.5-VL2026.06		48.39	49.52	49.36	47.36	50.54	47.04	44.8	48.14
Gemini-3.1-Flash2026.06		48.21	48.76	47.38	45.8	47.53	44.15	42.81	46.38
Qwen2.5-VL2026.06		46.86	46.2	49.47	44.99	49.67	44.7	43.98	46.55
Gemma-42026.06		44.2	43.29	43.46	42.47	44.49	42.4	40.12	42.92
GLM-4.6V-Flash2026.06		43.7	42.69	42.78	42.05	43.54	42.47	40.28	42.5
Grok-4.1-Fast2026.06		43.43	44.76	43.95	43.06	43.94	41.74	40.5	43.05
Claude-sonnet-4-62026.06		43.12	43.63	44.31	43	44.98	42.62	41.06	43.25
Claude-opus-4-62026.06		42.77	43.32	43.75	42.8	44.42	42.91	40.84	42.97