Egocentric Video Understanding on EgoSchema (test)
75.6AccuracyGPT-5
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5Category=API call only2026.03 | 75.6 | |
| Claude Sonnet 4.5Category=API call only2026.03 | 73.1 | |
| Gemini 2.5 ProCategory=API call only2026.03 | 72.2 | |
| Eagle2.5-8BCategory=Open weights2026.03 | 72.2 | |
| GPT-5 miniCategory=API call only2026.03 | 70.9 | |
| Gemini 2.5 FlashCategory=API call only2026.03 | 70.2 | |
| Qwen3-VL-8BCategory=Open weights2026.03 | 69.8 | |
| Gemini 3 ProCategory=API call only2026.03 | 68.9 | |
| PLM-8BCategory=Fully open2026.03 | 68.8 | |
| Qwen3-VL-4BCategory=Open weights2026.03 | 68.4 | |
| PLM-3BCategory=Fully open2026.03 | 66.9 | |
| MolmoPoint-8BCategory=MolmoPoint2026.03 | 63 | |
| GLM-4.1V-9BCategory=Open weights2026.03 | 62.6 | |
| Molmo2-8BCategory=Fully open2026.03 | 62 | |
| Molmo2-4BCategory=Fully open2026.03 | 61.2 | |
| InternVL3.5-4BCategory=Open weights2026.03 | 58.9 | |
| InternVL3.5-8BCategory=Open weights2026.03 | 58.6 | |
| LLaVA-Video-7BCategory=Fully open2026.03 | 57.3 | |
| Molmo2-O-7BCategory=Fully open2026.03 | 56.8 | |
| Keye-VL-1.5-8BCategory=Open weights2026.03 | 56.3 | |
| VideoChat-Flash-7BCategory=Fully open2026.03 | 51.3 | |
| TGPOBackbone=Qwen2.5-VL (3B)2026.03 | 49.7 | |
| MiniCPM-V-4.5-8BCategory=Open weights2026.03 | 49.6 | |
| EgoVLM GRPOParameters=3B2026.03 | 46.5 | |
| LLaVA-VideoParameters=72B2026.03 | 42.7 | |
| Llava-OneVisionParameters=72B2026.03 | 36.2 | |
| LLaVA-VideoParameters=7B2026.03 | 34.1 | |
| Qwen2.5-VLParameters=3B2026.03 | 20.4 |