Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multimodal Reasoning on MathVerse Vision Only
Loading...
47.1
Accuracy
LUSPO
40.86
42.48
44.1
45.72
Feb 5, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
LUSPO
Base model=Qwen2.5-VL-...
2026.02
47.1
GRPO
Base model=Qwen2.5-VL-...
2026.02
46.8
GSPO
Base model=Qwen2.5-VL-...
2026.02
45.6
Qwen2.5-VL-7B-Instruct (w/o RLVR)
Base model=Qwen2.5-VL-...
2026.02
41.1
Feedback
Search any
task
Search any
task
