Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
STEM Puzzle on MathVista
Loading...
83.9
Accuracy
CoT+M2Note
67.572
71.811
76.05
80.289
Jul 1, 2026
Accuracy
Updated 23d ago
Evaluation Results
Method
Method
Links
Accuracy
CoT+M2Note
Backbone=Qwen3-VL-8B-I...
2026.07
83.9
CoT
Backbone=Qwen3-VL-8B-I...
2026.07
82.9
M2Note
Backbone=Qwen3-VL-Plus...
2026.07
80.1
M2Note
Backbone (Tuning)=Qwen...
2026.07
78.3
Vanilla
Backbone=Qwen3-VL-Plus...
2026.07
78
M2Note
Backbone=GPT-5.4, Evol...
2026.07
78
M2Note
Backbone=Qwen3-VL-8B-I...
2026.07
77.3
Vanilla
Backbone=GPT-5.4, Evol...
2026.07
77.2
DPE
Backbone=Qwen3-VL-8B-I...
2026.07
76.2
Vanilla
Backbone=Qwen3-VL-8B-I...
2026.07
73.2
Vanilla
Backbone (Tuning)=Qwen...
2026.07
73.2
Vision-Zero
Evolving Mode=Self-Evo...
2026.07
72.6
EvoLMM
Evolving Mode=Self-Evo...
2026.07
70.5
iReasoner
Evolving Mode=Self-Evo...
2026.07
69.7
VisPlay
Evolving Mode=Self-Evo...
2026.07
68.2
Feedback
Search any
task
Search any
task