MOS
Benchmark website →Mean Opinion Score rates speech synthesis quality on a 1-5 scale, normalized to 0-100 for this platform.
About this test
- What it measures
- Perceived naturalness and quality of synthesized speech.
- How it was administered
- Human listeners rate samples 1-5; averaged and scaled to 0-100. Higher means more natural-sounding.
Model rankings
Models ranked by score on this benchmark. Higher is better.
| Rank | Model | Provider | Score | Percentile | Tags |
|---|---|---|---|---|---|
| 1 | ElevenLabs | 92.0 | p98 | Text-to-Speech, Proprietary | |
| 2 | OpenAI | 85.0 | p88 | Text-to-Speech, Proprietary | |
| 3 |  |
