E3mo-Bench (Evoked) - Single-Video Perception (What-How): leaderboard
Metric: Accuracy (%; what or how multiple-choice questions about one video's emotion, valence, arousal, dominance or its cause, one per video among 2,442 single-video questions; 16 MLLMs on videos from eight emotion datasets, each assigned one affective perspective; subset: the viewer's evoked emotion). Source: arxiv.org. Saturation forecast: Around 2030. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 66.18 |
| 2 | Qwen2.5-Omni-7B | 66.01 |
| 3 | Claude Sonnet 4.6 | 64.21 |
| 4 | InternVL3-8B | 60.85 |
Interactive version: theaggregate.ai/benchmark?slug=e3mo-bench-evoked-single-video-perception-what-how · How It Works · Data refreshed daily, snapshot 2026-09-29.