Video Understanding on EgoSchema fullset
72.2EgoSchema AccuracyGPT-4o
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-4o#Params.=-2026.06 | 72.2 | — | |
| VideoLLaMA3#Params.=7B2026.06 | 63.3 | — | |
| Gemini-1.5-Pro#Params.=-2026.06 | 63.2 | — | |
| LLaVA-OV#Params.=7B2026.06 | 60.1 | — | |
| LLaVA-Video#Params.=7B2026.06 | 57.3 | — | |
| GPT-4V#Params.=-2026.06 | 55.6 | — | |
| Vega#Params.=3B2026.06 | 52.9 | — | |
| VideoLLaMA2#Params.=7B2026.06 | 51.7 | — | |
| HaploOmni#Params.=9B2026.06 | 47.1 | — | |
| AVS + SVC2026.02 | — | 62.7 | |
| LLaVA-NeXT-Video-7BParameter Scale=7B2026.02 | — | 43.9 | |
| LLaVA-OneVision-7BParameter Scale=7B2026.02 | — | 60.1 | |
| LLoVi2026.02 | — | 52.2 | |
| VideoAgent2026.02 | — | 54.1 |