Fact · Dictation in your apps
10percent
OpenAI says word error rate in its ChatGPT dictation evaluations was at least 10% lower for the top languages tested than with the previous production model.
Checked against its source on October 17, 2026
In the source's own words
“was at least 10% lower for top languages tested than with the previous production model.”
ChatGPT release notes (OpenAI Help Center), OpenAI, OpenAI Help Center, 2026, June 26, 2026 entry.
Open the original- What it measures
- The release note describes word error rate as a standard measure of transcription accuracy.
- What it does not show
- A vendor's own evaluation. The note does not name the languages, the test set or the previous model, so outsiders cannot repeat it.