Visual Perception on CVBench 2D
79.09ScoreQwen2.5-VL
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-VLSize=72B2026.07 | 79.09 | |
| GPT-5.2Size=–2026.07 | 76.81 | |
| BRAIDSize=7B2026.07 | 76.26 | |
| GPT-4oSize=–2026.07 | 75.77 | |
| SFTSize=7B2026.07 | 74.85 | |
| BAGELSize=7B2026.07 | 73.53 | |
| Qwen3.5-VLScale=4B2026.05 | 72.4 | |
| Qwen3-VLScale=4B2026.05 | 72.1 | |
| DatologyScale=4B2026.05 | 71.6 | |
| Qwen2.5-VLSize=7B2026.07 | 68.25 | |
| DatologyScale=2B2026.05 | 65.2 | |
| Janus-ProSize=7B2026.07 | 63.82 | |
| InternVL3.5Scale=2B2026.05 | 63.6 | |
| InternVL3.5Scale=4B2026.05 | 62.9 | |
| Qwen3.5-VLScale=2B2026.05 | 62.9 | |
| InternVL3Scale=2B2026.05 | 60.8 | |
| Qwen3-VLScale=2B2026.05 | 60 | |
| DatologyScale=1B2026.05 | 59.8 | |
| MAmmoTH-VLScale=4B2026.05 | 53.5 | |
| MAmmoTH-VLScale=2B2026.05 | 47.2 | |
| MAmmoTH-VLScale=1B2026.05 | 30.6 | |
| ChameleonSize=7B2026.07 | 25.59 |