Ambiguous Speech Emotion - IEMOCAP: leaderboard
Ambiguous emotion recognition from speech: an audio-language model outputs a probability distribution over neutral, happy, angry and sad for 4,373 acted IEMOCAP utterances, compared with the spread of the human annotators' labels (Jia et al., 2026). Jensen-Shannon divergence between the two distributions, from 0 to 1; lower is better.
Metric: Jensen-Shannon divergence (0-1; between the emotion distribution the model predicts for an utterance and the distribution of the human annotators' labels, mean over utterances; lower is better; IEMOCAP, 4,373 acted dyadic utterances with neutral, happy, angry and sad labels). Source: arxiv.org. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 0.34 | #262 |
| 2 | Gemini 2.0 Flash | 0.35 | #702 |
| 3 | Qwen2.5-Omni-7B | 0.39 | #1797 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=ambiguous-speech-emotion-iemocap · How It Works · Data refreshed daily, snapshot 2026-11-12.