OpenAI Whisper (self-hosted)
Open-source speech-to-text model. Reported as the most accurate on clean audio, but needs technical setup and has no native speaker labels.
Pricing, as reported
Free when self-hosted (you supply the hardware).
Taken from third-party write-ups, not confirmed with the vendor. Checked Oct 27, 2026. Prices change often; confirm before you buy.
Where it ranks
| Use case | Rank | Score |
|---|---|---|
| Transcription | 2 of 4 | 75 |