Fact · Speech recognition
70times real time
WhisperX reports 70 times real-time transcription with the large-v2 model, using batched inference.
Checked against its source on October 17, 2026
In the source's own words
“This repository provides fast automatic speech recognition (70x realtime with large-v2) with word-level timestamps and speaker diarization.”
WhisperX: README (GitHub, m-bain/whisperX), m-bain/whisperX project, GitHub (m-bain/whisperX), 2026, Introduction.
Open the original- What it measures
- The README says WhisperX uses batched whisper with the faster-whisper backend, which is how the speed-up is obtained.
- What it does not show
- The figure is the project's own, measured with batched inference on a GPU. It measures speed, not accuracy.