EgoArgus (Assistance Decision) - Video-Grounded On-Topic: leaderboard
EgoArgus (Tang, Liu and Yen, 2026) evaluates vision-language models as egocentric assistants: from a first-person video, a dialogue and its timestamp, the model decides whether help is needed in 789 VISTA episodes. This board covers episodes where the video decides and the dialogue is related but uninformative. Assistance-decision F1 in percent; higher is better.
Metric: Assistance-decision F1 (%; whether help is required, from the egocentric video, dialogue and dialogue timestamp of 789 VISTA episodes, video-grounded on-topic scenario: related but uninformative dialogue; API models get the full video, open models sampled frames). Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Flash Lite | 80.49 | #363 |
| 2 | Molmo2-8B | 76.55 | #1488 |
| 3 | Qwen 3.5 Plus (2026-02-15) | 72.49 | #178 |
| 4 | Qwen 3.5 2B | 20.95 | #4849 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=egoargus-assistance-decision-video-grounded-on-topic · How It Works · Data refreshed daily, snapshot 2026-11-12.