Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Geometry Reasoning on Geometry3K (val)
Loading...
28.12
Accuracy
PEPO_D
7.0704
12.5352
18
23.4648
Mar 24, 2026
Accuracy
Updated 24d ago
Evaluation Results
Method
Method
Links
Accuracy
PEPO_D
Backbone=InternVL3-2B-...
2026.03
28.12
PEPO_G
Backbone=InternVL3-2B-...
2026.03
25.84
DAPO
Backbone=Qwen2.5-VL-3B...
2026.03
22.63
PEPO_D
Backbone=Qwen2.5-VL-3B...
2026.03
22.38
GRPO
Backbone=InternVL3-2B-...
2026.03
22.08
PEPO_G
Backbone=Qwen2.5-VL-3B...
2026.03
21.91
DAPO
Backbone=InternVL3-2B-...
2026.03
20.94
GRPO
Backbone=Qwen2.5-VL-3B...
2026.03
19
High-Entropy RL
Backbone=Qwen2.5-VL-3B...
2026.03
15.58
Base
Backbone=Qwen2.5-VL-3B...
2026.03
11.58
High-Entropy RL
Backbone=InternVL3-2B-...
2026.03
10.89
Base
Backbone=InternVL3-2B-...
2026.03
7.88
Feedback
Search any
task
Search any
task