EgoArgus (Assistance Decision) - Multimodal Grounded: leaderboard

Metric: Assistance-decision F1 (%; whether help is required, from the egocentric video, dialogue and dialogue timestamp of 789 VISTA episodes, multimodal-grounded scenario: dialogue and video are both needed; API models get the full video, open models sampled frames). Source: arxiv.org. Saturation forecast: Estimated already saturated. 7 models tracked.

Top models

#ModelScore
1Gemini 3.1 Flash Lite93.95
2Qwen 3.5 Plus (2026-02-15)93.4
3MiMo-V2-Omni91.94
4Molmo2-8B82.65
5Qwen 3.5 2B34.59

Interactive version: theaggregate.ai/benchmark?slug=egoargus-assistance-decision-multimodal-grounded · How It Works · Data refreshed daily, snapshot 2026-09-26.