Video Question Answering on EgoLife
45.4AccuracyTaskMem
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TaskMem2026.05 | 45.4 | 56.4 | 80.5 | |
| Gemini-2.5-Pro2026.05 | 43.8 | 56.6 | 77.4 | |
| Gemini-1.5-Pro2026.05 | 39.4 | 51.6 | 76.4 | |
| Qwen3-VL-30B-A3B2026.05 | 38.4 | 52.4 | 73.3 | |
| GPT-5.22026.05 | 34.8 | 48.2 | 72.2 | |
| HippoMM2026.05 | 30.4 | 43.4 | 70 | |
| M3-Agent2026.05 | 21.8 | 30.8 | 70.8 | |
| EgoGPT2026.05 | 19.2 | 28.2 | 68.1 |