EmoMM - Conflict: leaderboard
The Conflict subset of EmoMM (Sun et al., 2026): test clips in which at least one of text, audio or video conveys a sentiment polarity that conflicts with the overall label. Five-class sentiment accuracy in percent; higher is better.
Metric: Accuracy (%; EmoMM-Conflict test subset, at least one modality's polarity conflicts with the ground truth; five-class sentiment). Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 56.23 | #135 |
| 2 | Qwen2.5-Omni-7B | 46.47 | #1797 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=emomm-conflict · How It Works · Data refreshed daily, snapshot 2026-11-12.