Bench · Speech to text · October 2, 2026

Which voice model hears you?

Holly listens on your own devices. So we timed and scored the speech recognizers it can use, on English, Chinese, and both in one sentence.

errors for Whisper large v3, the most accurate on the Mac, across all three recordings
its typical wait per turn on a MacBook Air
errors for Apple's recognizer with 2 preset languages, English and Chinese, with a typical wait of

Speed and accuracy

Lower left is better: fewer errors, less waiting. Each dot is one recognizer, over turns of real speech.

Typical seconds per turn against average error rate 0%5%10%15%20% 0 s0.5 s1 s1.5 s Typical wait per turn (p50, seconds) Errors (average of three recordings) better

Accuracy

Errors per character, against what the speaker checked by ear. Lower is better. Every recognizer on the Mac heard the same turns.

*Apple's recognizer hears one language at a time. Buddy runs one for each of the 2 languages you set (here English and Chinese) together, and keeps whichever wrote more words in its own language. That's the row scored here.

†For comparison: a cloud model (GPT-5.6) asked for a word-for-word transcript of each whole recording. Not timed, and it has a head start: for English and mixed, the answer key began as its own earlier transcript, which the speaker then checked.

Speed

Time from the end of a turn to its words, on a MacBook Air. A typical turn was seconds long; the longest, .

*Here Apple was handed each finished turn at once, and the 2 languages count as the slower of the two. On a call it listens while you talk, so its wait after you stop should be shorter. We haven't measured that yet.

The models

How big each one is, and the exact copy we ran. Bigger and more precise usually means slower and more accurate.

Parameters are the model's learned numbers, as OpenAI publishes them. Precision is read from each model file. Palettized means each layer's numbers are snapped to a small shared set (16 values for 4-bit, 64 for 6-bit), which makes the file smaller: this turbo is 1.0 GB, where a 16-bit copy would be about 1.6 GB. The turbo we tested is that compressed copy, so some of its errors may come from compression; the full 16-bit turbo is untested. Apple doesn't publish its recognizer's size.

Listen and compare

Play a recording, then see exactly what each recognizer wrote, turn by turn, as it wrote it. The recordings here are turned up for listening; the recognizers heard the originals.

What we learned

A bug we found

Calls ask Whisper to work fast: no timestamps, a cap on words. Past 30 seconds Whisper needs timestamps to find where to keep reading, so a long turn lost most of its words. On a 65 second recording the biggest model kept 1 of 65 English words. The fix keeps timestamps on for anything over 30 seconds, and the same recording now comes through whole.

How we tested

What this test can't tell you

Every recognizer tested here ran on the Mac, except GPT-5.6, a cloud model we scored for comparison. Two of the three answer keys started from its transcripts.