Distil-Whisper Large v3
6x-faster English Whisper distillation within 1% word error rate.
About
Hugging Face's Distil-Whisper (distil-large-v3, 756M parameters) distills Whisper large-v3 for English speech recognition: 6.3x faster and 49% smaller, staying within about 1% word error rate on out-of-distribution sets — team-reported short-form WER of 9.7 versus 8.4 for the teacher. Apache 2.0 and native to Hugging Face Transformers. Honest caveats: English only — for other languages the Distil-Whisper team itself points to Whisper Turbo; and long-form WER, while close, still trails the teacher slightly.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Team-reported: 6.3x faster than large-v3; short-form WER 9.7 vs 8.4.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: distil-large-v3: 756M params, 6.3x faster, WER 9.7 vs 8.4 — team-reported figures verified via project README.; Apache 2.0; English-only limitation honestly disclosed with pointer to Whisper Turbo..