General Visual Question Answering on MMBench-EN (test)
92.05AccuracySTEP3-VL-10B
Evaluation results
| Method | Links | |
|---|---|---|
| STEP3-VL-10BNumber of Parameters=10B2026.01 | 92.05 | |
| GLM-4.6V FlashNumber of Parameters=9B2026.01 | 91.04 | |
| Qwen3-VL ThinkingNumber of Parameters=8B2026.01 | 90.55 | |
| MiMo-VL RL-2508Number of Parameters=7B2026.01 | 89.91 | |
| Qwen2.5-VLNumber of Parameters=72B2025.02 | 88.4 | |
| InternVL 3.5Number of Parameters=8B2026.01 | 88.2 | |
| InternVL2.5Number of Parameters=78B2025.02 | 87.4 | |
| Qwen2-VLNumber of Parameters=72B2025.02 | 86.1 | |
| GPT-4o 0513Model Version=05132025.02 | 83.1 | |
| Qwen2.5-VLNumber of Parameters=7B2025.02 | 82.6 | |
| Claude-3.5 Sonnet-0620Model Version=06202025.02 | 80.9 | |
| Qwen2.5-VLNumber of Parameters=3B2025.02 | 77.4 | |
| Qwen2.5-VLNumber of Parameters=72B2025.02 | 0.886 | |
| InternVL2.5Number of Parameters=78B2025.02 | 0.883 | |
| Qwen2-VLNumber of Parameters=72B2025.02 | 0.869 | |
| Qwen2.5-VLNumber of Parameters=7B2025.02 | 0.835 | |
| GPT-4o 0513Model Version=05132025.02 | 0.834 | |
| Claude-3.5 Sonnet-0620Model Version=06202025.02 | 0.826 | |
| Qwen2.5-VLNumber of Parameters=3B2025.02 | 0.791 |