Visual Question Answering on MMIU
71AccuracyGPT-5
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5Access=API call only2026.03 | 71 | |
| Gemini 2.5 ProAccess=API call only2026.03 | 68.9 | |
| GPT-5 miniAccess=API call only2026.03 | 64.5 | |
| GLM-4.1V-9BAccess=Open weights only2026.03 | 62.4 | |
| Gemini 2.5 FlashAccess=API call only2026.03 | 61.2 | |
| Molmo2-4BAccess=Fully open2026.03 | 55.5 | |
| MolmoPoint-8BAccess=MolmoPoint2026.03 | 54.6 | |
| Molmo2-8BAccess=Fully open2026.03 | 54.2 | |
| Claude Sonnet 4.5Access=API call only2026.03 | 54.1 | |
| MolmoPoint-8B-O-7BAccess=Fully open2026.03 | 51.7 | |
| Keye-VL-1.5-8BAccess=Open weights only2026.03 | 50.3 | |
| InternVL3.5-8BAccess=Open weights only2026.03 | 49.4 | |
| InternVL3.5-4BAccess=Open weights only2026.03 | 49.2 | |
| Eagle2.5-8BAccess=Open weights only2026.03 | 48.4 | |
| MiniCPM-V-4.5-8BAccess=Open weights only2026.03 | 46.5 | |
| Qwen3-VL-4BAccess=Open weights only2026.03 | 43.2 | |
| PLM-3BAccess=Fully open2026.03 | 40.6 | |
| Qwen3-VL-8BAccess=Open weights only2026.03 | 35.3 | |
| PLM-8BAccess=Fully open2026.03 | 27.4 |