← Speech-to-Text

Deepgram

Nova-3

Measured 2026-10-01

Run Nova-3 on Speko →

One API, every model on this board.

WER · batch

11.96%

n=50 · FLEURS clean read-English

Corrected 2026-09-02 from a published 9.8% that has no reproducible source. 11.96% is the mean per-clip WER of the committed batch artifact results/anchor-nova3-20260808.json (run of 2026-08-08), re-confirmed by rescoring that artifact with the current text normalizer — 11.96% as stored and 11.96% rescored.

WER · streaming

Read speech

10.8%

Conversation

16.95%

Details

Read speech

n=1541 · FLEURS test+dev+train · vendor-direct

Conversation

14.4–19.9% · EdAcc n=283 · 60 speakers, 2 passes, 2026-10-01

1 of 566 readings came back empty, scored as all-deletion.

Finalize

189ms

p50–p90 · US East · n=200 · vendor-direct · ±9%

Time to first token

976ms

p50–p90 · US East · n=200 · vendor-direct · ±0%

Cost / min

$0.0048

Streaming pay-as-you-go. Vendor flags it as a limited-time promotional rate; pre-recorded is $0.0043.

Share this result

Embed the live badge

Speko stt rank
Markdown
[![Speko stt rank](https://speko.ai/benchmarks/badge/stt/deepgram-nova-3.svg)](https://speko.ai/benchmarks/stt/deepgram-nova-3)
HTML
<a href="https://speko.ai/benchmarks/stt/deepgram-nova-3"><img src="https://speko.ai/benchmarks/badge/stt/deepgram-nova-3.svg" alt="Speko stt rank"></a>
URL
https://speko.ai/benchmarks/stt/deepgram-nova-3