SEABED - Affective Interpretation (Open-Ended): leaderboard
Metric: Accuracy (%; 504 joint valence and arousal judgment from prosody items; open-ended answers judged by Grok 4.5, binary correct or incorrect (graded 0 to 1 for long-form); audio at 16 kHz). Source: arxiv.org. Saturation forecast: Around August 2028. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.5 Flash | 37.7 |
| 2 | Gemini 3.6 Flash | 36.6 |
| 3 | Gemini 2.5 Pro | 30.8 |
| 4 | SeaLLMs-Audio-7B | 25 |
Interactive version: theaggregate.ai/benchmark?slug=seabed-affective-interpretation-open-ended · How It Works · Data refreshed daily, snapshot 2026-09-26.