List your company
ResearchFormerly WizWand

Multimodal Reasoning on MathVerse Vision Only

58.76AccuracyInternVL2.5-38B + VRPRM
21.361631.070840.7850.4892Aug 5, 2025Sep 4, 2025Oct 5, 2025Nov 5, 2025Dec 5, 2025Jan 5, 2026Feb 5, 2026
Updated 5mo ago

Evaluation results

MethodLinks
InternVL2.5-38B + VRPRMBase Model=InternVL2.5-38B, Selection Strategy=Best-of-8, Critic Model=VRPRM2025.0858.76
InternVL2.5-26B + VRPRMBase Model=InternVL2.5-26B, Selection Strategy=Best-of-8, Critic Model=VRPRM2025.0853.55
InternVL2.5-8B + VRPRMBase Model=InternVL2.5-8B, Selection Strategy=Best-of-8, Critic Model=VRPRM2025.0851.52
InternVL2.5-38B + VRPRM w/o RLBase Model=InternVL2.5-38B, Selection Strategy=Best-of-8, Critic Model=VRPRM w/o RL2025.0851.52
Gemini-2.0-FlashSelection Strategy=Standard2025.0847.8
LUSPOBase model=Qwen2.5-VL-7B-Instruct, Algorithm=LUSPO2026.0247.1
GRPOBase model=Qwen2.5-VL-7B-Instruct, Algorithm=GRPO2026.0246.8
InternVL2.5-38B + VisualPRMBase Model=InternVL2.5-38B, Selection Strategy=Best-of-8, Critic Model=VisualPRM2025.0846.7
Claude-3.5-SonnetSelection Strategy=Standard2025.0846.3
GSPOBase model=Qwen2.5-VL-7B-Instruct, Algorithm=GSPO2026.0245.6
InternVL2.5-26B + VRPRM w/o RLBase Model=InternVL2.5-26B, Selection Strategy=Best-of-8, Critic Model=VRPRM w/o RL2025.0844.29
Qwen2.5-VL-7B-Instruct (w/o RLVR)Base model=Qwen2.5-VL-7B-Instruct, Algorithm=w/o RLVR2026.0241.1
GPT-4oSelection Strategy=Standard2025.0840.6
InternVL2.5-8B + VRPRM w/o RLBase Model=InternVL2.5-8B, Selection Strategy=Best-of-8, Critic Model=VRPRM w/o RL2025.0839.85
InternVL2.5-26B + VisualPRMBase Model=InternVL2.5-26B, Selection Strategy=Best-of-8, Critic Model=VisualPRM2025.0839.1
InternVL2.5-38BBase Model=InternVL2.5-38B, Selection Strategy=Single-shot2025.0836.9
InternVL2.5-8B + VisualPRMBase Model=InternVL2.5-8B, Selection Strategy=Best-of-8, Critic Model=VisualPRM2025.0835.8
InternVL2.5-26BBase Model=InternVL2.5-26B, Selection Strategy=Single-shot2025.0824
InternVL2.5-8BBase Model=InternVL2.5-8B, Selection Strategy=Single-shot2025.0822.8