EgoExo-Next: leaderboard
Metric: Accuracy (%), unweighted mean over the eight subtask-by-scene settings (four-choice visual-option questions, chance 25%), zero-shot, mean of three runs. Source: arxiv.org. Saturation forecast: Around 2030. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 43.81 |
| 2 | GPT-5.2 | 43.16 |
| 3 | Qwen 3 VL 235B A22B | 40.93 |
| 4 | Claude Opus 4.6 | 33.59 |
Interactive version: theaggregate.ai/benchmark?slug=egoexo-next · How It Works · Data refreshed daily, snapshot 2026-10-07.