whatmodel.sh / stt / Rev AI Machine (Reverb ASR)

Rev AI Machine (Reverb ASR)

+ contribute

Rev AI Machine (Reverb ASR) is a closed-weights speech-to-text model from Rev. Licence: proprietary.

Catalogue identity notes

Rev AI async API transcriber='machine' (leaderboard revai_provider.py), which Rev documents as its Reverb ASR model. An unversioned service tier, like catalogue speechmatics-enhanced; release month unknown.

52.0±3.3#89 of 162 speech-to-text modelsfrom 3 benchmark results across 2 sources

Score within the Open ASR English short-form WER (March 2026) comparison set only.

$ how it compares

The speech-to-text models ranked either side of Rev AI Machine (Reverb ASR), out of 176 with a score, then more from Rev.

#	model	score	where	RAM	released
98	Omnilingual ASR CTC 7B v2 Meta	52±3	local	~16GB	2025-12
99	Omnilingual ASR LLM 300M Meta	52±2	local	~4.4GB	2025-11
100	Rev AI Machine (Reverb ASR) Rev	52±3	api $0.0033/audio min
101	LiteWhisper large-v3-fast Efficient-Speech (LiteASR)	52±3	local	~3.6GB	2025-02
102	VibeVoice-ASR-HF Microsoft	52±3	local	~21GB	2026-03
	more from Rev
70	Rev AI Fusion (Whisper Fusion) Rev	57±4	comparison only		2024-01

see all speech-to-text models · find the ones that fit your machine

$ what it's good at

long-form transcription▁▁▄▁·▁▂█▄▁▁▁ #7 of 27 * · 37% of its score · 1 test

Beats Chirp 3, Parakeet TDT 0.6B v2, Parakeet TDT 1.1B here, though they rank higher overall.

test	result	vs the others in its run	difficulty	counts	why
Open ASR English long-form WER	8.8% ✎	▁▁▅▁·▃▅█▆▁▂▂ beats 21/30	avg 12%	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 37%	typical

other tests2 not in a task · 59% of its score

test	result	vs the others in its run	difficulty	counts	why
Open ASR English short-form WER (March 2026)	8.0% ✎	▁▁▁▃··▁▁▂██▆ beats 36/86	avg 8.2%	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 37%	blunt
verli STT benchmark · LibriSpeech test-clean WER	4.4% ✎	▃·····▃▅█▃▅▅ beats 0/11	avg 3.4%	▇▇▇▇▇▇▇▇▇ 22%	typical

$ where to run it

Through an API

api · $0.0033/audio min · docs.rev.ai ↗ · price sourceUSD per input audio minute: Rev AI lists 'Reverb Transcription' (English) at $0.20 per hour (÷60). Rev's transcriber docs state that transcriber=machine (the default) is performed by the Reverb ASR model. Product: Rev AI asynchronous speech-to-text. API model: transcriber=machine (Reverb ASR). Settings: English, pay-as-you-go. Price basis: published list rate, per-hour to per-minute scaling only; billed rounded up to the nearest second with a 15-second minimum. Exclusions: Reverb Foreign Language ($0.30/hr), Whisper Large, human transcription, free credits. Currency: USD. Rev AI official pricing page, observed 2026-10-07.

$ how the score was worked out · Open ASR Leaderboard 74%, verliapp/stt-benchmark · inde… 22%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

source	results	share of the score	licence
Open ASR Leaderboard	2	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 74%	Apache-2.0
verliapp/stt-benchmark · independent STT accuracy benchmark	1	▇▇▇▇▇▇▇ 22%	MIT
prior · pull toward the middle		▇ 4%
weighted average		= 52.0

$ improve this page

Update model information Add a place to run it Add a test result Propose a benchmark Correct a benchmark score Edit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated 2026-10-11 02:20 UTC · sources and licences · json