Multimodal Reasoning on MathVerse VO
48.71FEI ScoreMiMo-VL-7B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MiMo-VL-7B# Samples=unk2025.08 | 48.71 | 66.15 | |
| VRPRM-MiMo w/o RL# Samples=3.6K, Explicit Reasoning (CoT)=true, RL Training=false2025.08 | 48.46 | 61.87 | |
| VRPRM-MiMo# Samples=53.6K, Explicit Reasoning (CoT)=true, RL Training=true2025.08 | 47.95 | 71.93 | |
| Qwen2.5-VL-7B# Samples=unk2025.08 | 46.69 | 42.75 | |
| VRPRM-MiMo w/o RL & w/o CoT# Samples=3.6K, Explicit Reasoning (CoT)=false, RL Training=false2025.08 | 46.69 | 46.55 | |
| VRPRM-Qwen3 w/o RL# Samples=3.6K, Explicit Reasoning (CoT)=true, RL Training=false2025.08 | 46.49 | 57.46 | |
| VRPRM-Qwen3 w/o RL & w/o CoT# Samples=3.6K, Explicit Reasoning (CoT)=false, RL Training=false2025.08 | 46.39 | 53.72 | |
| Qwen3-VL-4B-Thinking# Samples=unk2025.08 | 46.2 | 63.86 | |
| VRPRM-MiMo w/o CoT# Samples=53.6K, Explicit Reasoning (CoT)=false, RL Training=true2025.08 | 45.42 | 71.05 | |
| VRPRM-Qwen3# Samples=53.6K, Explicit Reasoning (CoT)=true, RL Training=true2025.08 | 44.93 | 70.19 | |
| VRPRM w/o RL# Samples=3.6K, Explicit Reasoning (CoT)=true, RL Training=false2025.08 | 43.96 | 62.43 | |
| InternVL2.5-8B# Samples=unk2025.08 | 42.96 | 43.62 | |
| Qwen2.5-VL-72B# Samples=unk2025.08 | 42.5 | 45.92 | |
| VRPRM w/o RL & w/o CoT# Samples=3.6K, Explicit Reasoning (CoT)=false, RL Training=false2025.08 | 41.72 | 51.7 | |
| VRPRM# Samples=53.6K, Explicit Reasoning (CoT)=true, RL Training=true2025.08 | 40.94 | 63.8 | |
| VRPRM-Qwen3 w/o CoT# Samples=53.6K, Explicit Reasoning (CoT)=false, RL Training=true2025.08 | 38.99 | 67.23 | |
| Gemini-2.0-Flash# Samples=unk2025.08 | 36.16 | 40.89 | |
| VRPRM w/o CoT# Samples=53.6K, Explicit Reasoning (CoT)=false, RL Training=true2025.08 | 34.8 | 54.72 | |
| GPT-4o-mini# Samples=unk2025.08 | 28.36 | 34.44 | |
| VisualPRM-8B# Samples=400K2025.08 | 24.56 | 60.93 |