Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Mathematical & Geometric Reasoning on DynaMath (accuracy@8)
Loading...
73.1
Accuracy@8
Qwen2.5-VL-32B + VPPO
55.004
59.702
64.4
69.098
Oct 10, 2025
Accuracy@8
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy@8
Qwen2.5-VL-32B + VPPO
Model Scale=32B, Algor...
2025.10
73.1
Qwen2.5-VL-32B + DAPO
Model Scale=32B, Algor...
2025.10
72.6
NoisyRollout-32B
Model Scale=32B, Promp...
2025.10
72.2
MM-Eureka-32B
Model Scale=32B, Promp...
2025.10
72
Qwen2.5-VL-32B + GRPO
Model Scale=32B, Algor...
2025.10
71.6
Qwen2.5-VL-32B
Model Scale=32B
2025.10
68.7
Qwen2.5-VL-7B + VPPO
Model Scale=7B, Algori...
2025.10
68.1
PAPO-D-7B
Model Scale=7B, Traini...
2025.10
66.8
Qwen2.5-VL-7B + DAPO
Model Scale=7B, Algori...
2025.10
66.6
Qwen2.5-VL-7B + GRPO
Model Scale=7B, Algori...
2025.10
65.8
VL-Rethinker-7B
Model Scale=7B, Traini...
2025.10
65.7
NoisyRollout-7B
Model Scale=7B, Traini...
2025.10
65.5
MM-Eureka-7B
Model Scale=7B, Traini...
2025.10
65.4
R1-ShareVL-7B
Model Scale=7B, Traini...
2025.10
65.1
ThinkLite-7B
Model Scale=7B, Traini...
2025.10
64.6
Qwen2.5-VL-7B
Model Scale=7B
2025.10
55.7
Feedback
Search any
task
Search any
task