In our October 2026 test on a Mac with an M2 Pro, Whisper large-v3-turbo had a word error rate of 3.92 percent on 100 English FLEURS recordings as recorded (16 kHz, no Bluetooth codec).

Measured by us on October 22, 2026, from 100 recordings

Our own measurement

This number comes from our benchmark, not from another publication. The full record (the setup, every recording's checksum, every transcript and timing) is published, so you can recompute it.

Bluetooth Headset Dictation Accuracy, Tested Download the full record (JSON)

What it measures: OpenAI Whisper large-v3-turbo, 5-bit quantized (ggml-large-v3-turbo-q5_0.bin), run through whisper.cpp 1.9.4 with Metal on the GPU, with the language given. The same 100 English recordings in every condition, so the conditions differ only in the audio path.
How it is measured: Word error rate = 84 word errors / 2,141 reference words = 3.92 percent. Every recording was brought to the same speech level (-26 dB below full scale) and, for this condition, left as it was.
What it does not show: The codecs are run in software on studio-clean read speech. A real headset adds its own microphone, noise suppression and radio losses, which this test leaves out, so compare conditions and engines rather than this number with a headset in a room.

In the data

A row of Bluetooth Headset Dictation Accuracy, Tested, as Whisper large-v3-turbo, Original recording: 3.92 percent.

Engine: Whisper large-v3-turbo
Maker: OpenAI
Audio path: Original recording
Language: English

Used to calculate

The speech recognition accuracy through Bluetooth headset audio dataset holds 12 data points, each a published figure traced to its source.

Our own measurement

In the data

Used to calculate

Cited on