Results
Whisper Medium accuracy and speed
Whisper Medium trades accuracy for a smaller download and less memory. On an older Mac it is often the largest Whisper that runs comfortably.
60
Blended score
15.8%
Word error rate
0.93 s
Wait for the text
Macro average over 8 datasets on Apple M4
On Apple M4
| Dataset | Score ↑ | WER ↓ | Speed ↑ | Recall ↑ | F-score ↑ | Response ↓ |
|---|---|---|---|---|---|---|
| AMI Meetingssw-app-local-sw-app-whisper-medium-esb-ami-20260806 | 59 | 16.2% | 13× | — | — | 0.93 s |
| Common Voice, Australian Englishsw-app-local-sw-app-whisper-medium-common-voice-24-en-au-20260806 | 79 | 8.4% | 6.8× | — | — | 0.93 s |
| Common Voice, spontaneous Englishsw-app-local-sw-app-whisper-medium-common-voice-spontaneous-4-en-20260806 | 54 | 18.4% | 11× | — | — | 0.93 s |
| Earnings 22sw-app-local-sw-app-whisper-medium-esb-earnings22-20260806 | 59 | 16.2% | 16× | — | — | 0.93 s |
| Earnings 22 (with vocabulary)sw-app-local-sw-app-whisper-medium-contextual-earnings22-20260806 | 25 | 40.3% | 14× | 59.6% | 72.7% | 0.93 s |
| LibriSpeech Othersw-app-local-sw-app-whisper-medium-esb-librispeech-other-20260806 | 82 | 7.1% | 13× | — | — | 0.93 s |
| Loquacioussw-app-local-sw-app-whisper-medium-loquacious-test-20260806 | 77 | 9.2% | 9.9× | — | — | 0.93 s |
| Phonetic vocabularysw-app-local-sw-app-whisper-medium-phonetic-vocab-20260806 | 73 | 10.7% | 4.0× | 62.9% | 69.8% | 0.93 s |
Updated August 27, 2026 from bench commit 5344cab. The second line of each row is the run that produced it.
Method
How this was measured
Local results come from driving the real Superwhisper app over each clip, so they reflect the engine and decode settings that ship in the product rather than a separate implementation.
Word error rate counts insertions, deletions and substitutions against a human transcript, so lower is better. Speed is a multiple of real time, so higher is better. The full method and every other model is on the benchmarks page.
Compare
Other models in the benchmark
Same audio, same machines, same method.
Support