Fact · Dictation in your apps
10times the relative speed of large
The Whisper README gives the tiny model a relative speed of about 10 times the large model, which it rates 1x.
Checked against its source on October 17, 2026
In the source's own words
“tiny 39 M tiny.en tiny ~1 GB ~10x base 74 M”
openai/whisper: Robust Speech Recognition via Large-Scale Weak Supervision, OpenAI, GitHub (OpenAI repository), 2023, Available models and languages table.
Open the original- What it measures
- Relative speed is the table's own comparison, with the large model as the baseline at 1x.
- What it does not show
- Relative speed depends on hardware and on the tool that runs the model. Treat it as a ratio, not as seconds of audio.