List your company

Multimodal Reasoning on MathVerse (mean@8 acc %)

60.95Mean@8 AccuracyDAPO
8.22221.91135.649.289May 10, 2026

Evaluation results

MethodLinks
DAPOBase Model=Qwen2-VL-8B Instruct, Temperature=0.62026.0560.95
PAPO_DBase Model=Qwen2-VL-8B Instruct, Temperature=0.62026.0559.64
VPPO_DBase Model=Qwen2-VL-8B Instruct, Temperature=0.62026.0559.64
RAPO_DBase Model=Qwen2-VL-8B Instruct, Temperature=0.62026.0558.5
RAPOBase Model=Qwen3-VL-8B-Instruct2026.0556.72
VPPOBase Model=Qwen3-VL-8B-Instruct2026.0556.35
GRPOBase Model=Qwen3-VL-8B-Instruct2026.0555.23
PAPOBase Model=Qwen3-VL-8B-Instruct2026.0554.57
BaseBase Model=Qwen3-VL-8B-Instruct2026.0550.21
Base (Qwen2-VL-8B Instruct)Temperature=0.62026.0550.21
RAPO_DBase Model=Qwen2.5-VL-7B-Instruct2026.0547.59
RAPO_GBase Model=Qwen2.5-VL-7B-Instruct2026.0547.34
R1-ShareVLBase Model=Qwen2.5-VL-7B-Instruct2026.0546.22
RAPOBase Model=Qwen3-VL-2B-Instruct2026.0546.13
VPPOBase Model=Qwen3-VL-2B-Instruct2026.0545.43
VL-RethinkerBase Model=Qwen2.5-VL-7B-Instruct2026.0545.3
PAPOBase Model=Qwen3-VL-2B-Instruct2026.0545.18
MM-EurekaBase Model=Qwen2.5-VL-7B-Instruct2026.0545.18
GRPOBase Model=Qwen3-VL-2B-Instruct2026.0544.35
BaseBase Model=Qwen3-VL-2B-Instruct2026.0536.75
RAPO_DBase Model=Qwen2-VL-7B-Instruct2026.0536.31
RAPO_GBase Model=Qwen2-VL-7B-Instruct2026.0532.99
BaseBase Model=Qwen2.5-VL-7B-Instruct2026.0529.44
MINT-CoTBase Model=Qwen2-VL-7B-Instruct2026.0524.62
TVCBase Model=Qwen2-VL-7B-Instruct2026.0521.98
BaseBase Model=Qwen2-VL-7B-Instruct2026.0510.25