Mathematical Reasoning on MathVista (GEO, ARI, GPS, ALL)
64.1Overall Score (ALL)R1-Onevision-7B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| R1-Onevision-7BData Amount=10k2025.07 | 64.1 | — | — | — | |
| GPT-4oData Amount=-2025.07 | 63.8 | — | — | — | |
| R1-VL-7BData Amount=10k (SFT 120k)2025.07 | 63.6 | — | — | — | |
| MGT-PerceReasonData Amount=65k2025.07 | 63.2 | — | — | — | |
| SyncLoop-3rdBackbone=Qwen2-VL-7B, Data Amount=0.4k2025.07 | 63.2 | 59 | 59 | 60 | |
| SyncLoop-2ndBackbone=Qwen2-VL-7B, Data Amount=0.6k2025.07 | 62.4 | 56 | 58 | 56 | |
| SyncLoop-1stBackbone=Qwen2-VL-7B, Data Amount=0.6k2025.07 | 62.1 | 55 | 57 | 54 | |
| Qwen2-VL-7BData Amount=-2025.07 | 60 | 50 | 50 | 49 | |
| LLamaV-o1-11BData Amount=>100k2025.07 | 54.4 | — | — | — | |
| SyncLoop-3rdBackbone=Qwen2-VL-2B, Data Amount=0.4k2025.07 | 50.2 | 40 | 40 | 38 | |
| Insight-V-8BData Amount=200K2025.07 | 49.8 | — | — | — | |
| SyncLoop-2ndBackbone=Qwen2-VL-2B, Data Amount=0.6k2025.07 | 49.3 | 38 | 38 | 38 | |
| SyncLoop-1stBackbone=Qwen2-VL-2B, Data Amount=0.6k2025.07 | 49.1 | 38 | 40 | 38 | |
| Qwen2-VL-2BData Amount=-2025.07 | 43 | — | — | — |