Category
Benchmark
Qwen2.5
HuatuoGPT
MedVLThinker
InternVL3.5
GLM-4.1V
QoQ-Med
LingShu
OctoMed
MedGemma
GPT-4o
DeepSeek-R1
Information
BaseLLM
Method(s)
SFTSFTSFT+RLSFT+RLSFT+RLRLSFT+RLSFT
SFT+RL-SFT+RL
ModelSize
7B7B7B8B9B7B7B7B
27B
>
100B671B
Text-Only
MedQA
57.9955.4058.5673.3776.8358.3062.09
90.81
†
85.1790.7293.16
HeadQA
69.6167.1471.0081.0081.7370.9469.23
82.36
84.5689.5690.93
MedMCQA
51.4550.7454.8760.7264.6453.1153.16
72.70
70.3477.2179.36
MedXpertQA
12.1610.8512.9112.8217.6711.7212.54
24.51
21.8930.3137.30
MMLU-PRO(H)
49.3244.7752.1362.9668.3349.9350.73
68.75
70.8376.0679.39
Overall
48.10
45.78
49.89
58.17
61.84
48.80
49.55
67.83
66.56
72.77
76.05
MultimodalReasoning
MMMU-PRO(H)
32.8730.0038.9547.90
49.65
33.2936.7842.52
40.7757.76-
NEJM
43.7644.0644.3152.9054.8144.0148.43
61.14
57.0669.99-
PMC-VQA
49.6651.4649.8859.5557.0549.7958.76
61.13
46.0860.14-
MedXpertQA
22.4722.6324.8126.1028.6522.7426.19
36.65
33.1344.38-
Overall
37.19
37.04
39.49
46.61
47.54
37.46
42.54
50.36
44.25
58.07
-
MultimodalClassification
BrainTumor
27.8237.1631.0764.4747.7255.3378.99
80.86
58.3365.99-
CoronaHack
34.1742.7636.4138.6239.9050.1342.24
71.22
58.2741.70-
Aptos
26.7140.9525.0851.9642.7664.6056.83
75.43
49.2066.58-
MESSIDOR2
20.6333.6622.4027.7126.86
59.14
42.8655.20
53.0945.26-
BCSS
43.1243.7248.0150.9552.8452.0252.40
53.72
35.9550.28-
Overall
30.49
46.39
32.59
46.39
41.76
56.24
54.66
67.29
50.97
53.96
-