Rankings
Prototype leaderboards for STT and TTS. The data is seeded for product validation and designed to become live evals.
Medical STT benchmark v4
42 models ranked by Medical WER (clinically relevant words only)
New metric: M-WER + Drug M-WERTop model: WhisperOpen-source leaderboard + eval code
* Updated 2026-05-06 • seeded from provided benchmark direction for prototype charting.
Top TTS models
Ranked by quality, latency, language coverage, cloning support, and cost curve.