Visual Understanding on CV-Bench
86.96AccuracyERNIE 5.0-Base
Evaluation Results
| Method | Links | |
|---|---|---|
| ERNIE 5.0-BaseModel type=pre-trained2026.02 | 86.96 | |
| InternVL3.5Size=38B2025.10 | 85.96 | |
| GPT-5Size=-2025.10 | 85.46 | |
| Gemini 2.5 FlashSize=-2025.10 | 85.07 | |
| Qwen2.5-VLSize=72B2025.10 | 82.54 | |
| InternVL3.5Size=8B2025.10 | 81.99 | |
| ThinkMorphSize=7B, think mode=true2025.10 | 80.82 | |
| CoLVR2026.05 | 76.95 | |
| BagelSize=7B, think mode=false2025.10 | 76.03 | |
| GPT-4oSize=-2025.10 | 75.61 | |
| LVRSteps=82026.05 | 75.51 | |
| Qwen2.5-VLSize=7B2025.10 | 75.2 | |
| Mirage2026.05 | 72.4 | |
| Janus-proSize=7B2025.10 | 67.83 | |
| ChameleonSize=7B2025.10 | 36.52 |