CBT-Audio (Audio + Transcript): leaderboard
Metric: Mean absolute error on the 1-5 distress-intensity scale against SSR reference labels (1,802 patient turns; lower is better); patient audio, its transcript and text context. Source: arxiv.org. Saturation forecast: Around May 2027. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | gemma-4-E4B-it | 0.72 |
| 2 | Qwen2.5-Omni-7B | 0.93 |
| 3 | Voxtral-Mini-3B-2507 | 1.1 |
| 4 | SALMONN-7B | 1.6 |
Interactive version: theaggregate.ai/benchmark?slug=cbt-audio-audio-plus-transcript · How It Works · Data refreshed daily, snapshot 2026-09-26.