ESCUCHA - Pathological Speech: leaderboard
Metric: Accuracy (%; 158 multiple-choice questions on non-normative speech from speakers with ALS or after a stroke; in-the-wild Spanish audio from 6.7 s to 85 min per question with multiple accents and non-normative speech; one shared system prompt for every model). Source: arxiv.org. Saturation forecast: Around 2032. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Voxtral-Mini-3B-2507 | 51.9 |
| 2 | Qwen2.5-Omni-7B | 44.94 |
| 3 | Gemini 2.5 Flash | 36.71 |
| 4 | Gemma 4 12B (IT) | 36.08 |
Interactive version: theaggregate.ai/benchmark?slug=escucha-pathological-speech · How It Works · Data refreshed daily, snapshot 2026-09-29.