Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Geometric Reasoning on Geometry3K
Loading...
43.8
Accuracy@1
VisuoThink
13.744
21.547
29.35
37.153
Apr 12, 2025
Accuracy@1
Updated 4d ago
Evaluation Results
Method
Method
Links
Accuracy@1
VisuoThink
Backbone=Claude-3.5-so...
2025.04
43.8
VisualSketchpad + Equation Solver
Backbone=Claude-3.5-so...
2025.04
41.7
VisualSketchpad
Backbone=Claude-3.5-so...
2025.04
39.6
CoT
Backbone=Claude-3.5-so...
2025.04
37.5
VisuoThink w/o rollout search
Backbone=Claude-3.5-so...
2025.04
37.5
VisuoThink
Backbone=GPT-4o
2025.04
33.3
VisuoThink w/o rollout search
Backbone=GPT-4o
2025.04
27.1
VisualSketchpad + Equation Solver
Backbone=GPT-4o
2025.04
25
VisuoThink
Backbone=Qwen2-VL-72B-...
2025.04
25
VisualSketchPad
Backbone=GPT-4o
2025.04
22.9
CoT
Backbone=GPT-4o
2025.04
20.8
VisuoThink w/o rollout search
Backbone=Qwen2-VL-72B-...
2025.04
20.8
CoT
Backbone=Qwen2-VL-72B-...
2025.04
18.8
VisualSketchpad
Backbone=Qwen2-VL-72B-...
2025.04
17
VisualSketchpad + Equation Solver
Backbone=Qwen2-VL-72B-...
2025.04
14.9
Feedback
Search any
task
Search any
task