Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Financial Multimodal Reasoning on FinMME
Loading...
49.64
Accuracy
Qwen2.5-VL
15.0808
24.0529
33.025
41.9971
Dec 9, 2025
Jan 2, 2026
Jan 26, 2026
Feb 19, 2026
Mar 15, 2026
Apr 8, 2026
May 2, 2026
Accuracy
Updated 2mo ago
Evaluation Results
Method
Method
Links
Accuracy
Qwen2.5-VL
Parameter Scale=72B
2025.12
49.64
Qwen2-VL
Parameter Scale=72B
2025.12
37.11
SYNDATA
Model=Qwen3-VL-4B, Dat...
2026.05
36.96
DeepSeekVL-2 Small
2025.12
34.14
Qwen2-VL
Parameter Scale=7B
2025.12
34.14
SYNDATA
Model=Qwen3-VL-4B, Dat...
2026.05
33.81
SYNDATA
Model=Qwen3-VL-4B, Dat...
2026.05
33.46
Qwen2.5 VL-7B + PHM
Compression Method=PHM...
2025.12
33.1
Qwen2.5-VL
Parameter Scale=3B
2025.12
32.53
SYNDATA
Model=Qwen2.5-VL-7B, D...
2026.05
32.47
LLaVA-1.5-7B + PHM
Compression Method=PHM...
2025.12
28.5
Base Model
Model=Qwen3-VL-4B, Dat...
2026.05
27.65
SYNDATA
Model=Qwen2.5-VL-7B, D...
2026.05
26.98
InstructBLIP
Backbone=Vicuna-13B
2025.12
26.21
SYNDATA
Model=Qwen2.5-VL-7B, D...
2026.05
24.9
InstructBLIP
Backbone=Vicuna-7B
2025.12
24.5
InstructBLIP + PHM
Compression Method=PHM...
2025.12
21.5
Base Model
Model=Qwen2.5-VL-7B, D...
2026.05
16.41
Feedback
Search any
task
Search any
task