To help you evaluate speech-to-text performance for your target language, please refer to the official supported languages list here:
https://platform.openai.com/docs/guides/speech-to-text#supported-languages

You may also review the attached accuracy chart for detailed error-rate comparisons across languages.

While our speech-to-text system (powered by OpenAI technology) supports many languages, transcription accuracy varies significantly depending on the language—particularly for less commonly used (lower-resource) languages.

For example:

• Widely spoken languages such as Dutch (4.3% error rate) and Spanish (4.7%) demonstrate very low error rates, meaning high transcription accuracy.

• Smaller or lower-resource languages (such as Albanian or Nepali) may show significantly higher error rates, in some cases exceeding 40%.

Before making a final decision, we recommend reviewing the error rate of your target language to ensure it meets your accuracy expectations. Generally, lower-resource languages will have comparatively lower recognition performance.

If you would like assistance evaluating a specific language, feel free to let us know.