Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Geometric Reasoning on GeoQA
Loading...
72.52
Accuracy
GeoSketch-Qwen2.5VL-7B
21.2688
34.5744
47.88
61.1856
Sep 26, 2025
Nov 2, 2025
Dec 10, 2025
Jan 17, 2026
Feb 23, 2026
Apr 2, 2026
May 10, 2026
Accuracy
Updated 22d ago
Evaluation Results
Method
Method
Links
Accuracy
GeoSketch-Qwen2.5VL-7B
Training Stage=RL
2025.09
72.52
Theorem-SFT
Backbone=Qwen2.5-VL-7B...
2026.05
69.73
Theorem-SFT
Backbone=Qwen2.5-VL-7B...
2026.05
66.89
GeoSketch-Qwen2.5VL-7B
Training Stage=SFT
2025.09
65.09
Base
Backbone=Qwen2.5-VL-7B
2026.05
49.46
Qwen2.5VL-7B
Mode=Base
2025.09
47.28
Theorem-SFT
Backbone=Qwen2.5-VL-3B...
2026.05
41.89
Vanilla SFT
Backbone=Qwen2.5-VL-7B
2026.05
38.78
Theorem-SFT
Backbone=Gemma-3-4B-IT...
2026.05
36.72
Base
Backbone=Qwen2.5-VL-3B
2026.05
35.95
Theorem-SFT
Backbone=Gemma-3-4B-IT...
2026.05
35.72
Theorem-SFT
Backbone=Qwen2.5-VL-3B...
2026.05
35.54
Vanilla SFT
Backbone=Qwen2.5-VL-3B
2026.05
34.59
Vanilla SFT
Backbone=Gemma-3-4B-IT
2026.05
31.89
Base
Backbone=Gemma-3-4B-IT
2026.05
23.24
Feedback
Search any
task
Search any
task