Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Science Diagram Reasoning on AI2D
Loading...
79.92
Accuracy
Coreset
-3.1968
18.3816
39.96
61.5384
Oct 24, 2025
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
Coreset
Backbone=Qwen2.5-VL-7B...
2025.10
79.92
RECAP
Backbone=Qwen2.5-VL-7B...
2025.10
78.21
Uniform
Backbone=Qwen2.5-VL-7B...
2025.10
76.43
VLAA-Thinker-7B
2025.10
75.45
PropMix
Backbone=Qwen2.5-VL-7B...
2025.10
75.32
Reasoning-only
Backbone=Qwen2.5-VL-7B...
2025.10
74.97
LwF
Backbone=Qwen2.5-VL-7B...
2025.10
73.93
Qwen2.5-VL-7B
variant=Base model
2025.10
67.62
OpenVLThinker-7B
2025.10
59.94
Vision-R1-7B
2025.10
0
Feedback
Search any
task
Search any
task