WER (inverted)
Benchmark website →Word Error Rate measures speech recognition accuracy. Shown here as accuracy (100 - WER) so higher is better.
About this test
- What it measures
- Speech-to-text transcription accuracy across diverse audio conditions.
- How it was administered
- Models transcribe audio; WER computed as (substitutions + insertions + deletions) / total words; inverted to accuracy for ranking.
Model rankings
Models ranked by score on this benchmark. Higher is better.
| Rank | Model | Provider | Score | Percentile | Tags |
|---|---|---|---|---|---|
| 1 | OpenAI | 94.5 | p97 | Speech Recognition, Open Source, Large |
