OpenTranscription
OpenTranscription
RankerModelsPlayground
OpenTranscription
OpenTranscription

One API to every speech-to-text model worth using. Compare them on your audio, route to the best one, pay per second.

Platform status

Product

RankerModelsTranscriptionsPlaygroundBlog

Developers

DocumentationReliabilityAPI VersioningStatus

Legal

Privacy PolicyTerms of ServiceSupport
© 2026 OpenTranscription
Back to Models

Nova-3 vs Scribe v2

2 models compared on the same benchmark — accuracy, latency, price, language coverage and capabilities. Scribe v2 leads on raw accuracy; Scribe v2 is fastest; Scribe v2 covers the most languages; Scribe v2 is cheapest.

Open both in playground

Verdict · who wins what

Most accurate

Scribe v2

9.2% WER

ElevenLabs

Fastest

Scribe v2

0.2× realtime

ElevenLabs

Most languages

Scribe v2

76 languages

ElevenLabs

Cheapest

Scribe v2

$0.0040/min

ElevenLabs

Comparing

2 models · same benchmark

1

Batch & Realtime

Nova-3

Deepgram

$0.0077/min
#29 · 71.1

Deepgram's flagship model — 53% lower WER vs competitors, code-switching support

Use This ModelDetails →

2

Batch

Scribe v2

ElevenLabs

$0.0040/min
#2 · 78.2

State-of-the-art batch STT — 90+ languages, speaker diarization, audio tagging

Use This ModelDetails →

Overview

30-day benchmark average

Overall score

0–100

71.1

78.2

Accuracy rank

of 35

#29

#2

Word error rate

WER

21.0%

9.2%

Character error rate

CER

11.5%

7.2%

Match error rate

MER

20.0%

8.5%

Word info lost

WIL

25.8%

12.3%

Languages

47

76

Cost vs. accuracy

overall WER vs price

Nova-3

Scribe v2

field · 33 models

lower-left is better

Accuracy by category

WER · lower is better

code_switching

63.0%
14.6%

conversational

12.2%
7.6%

finance

9.5%
7.6%

Accuracy by accent

WER · English

african

26.2%
21.5%

australian

15.4%
3.1%

indian

27.0%
9.5%

Speed & latency

p50 → p99

Processing speed

× realtime

0.1×

0.2×

Latency p50

1.3s

3.4s

Latency p90

1.8s

4.0s

Realtime / streaming

Streaming benchmark averages

Realtime rank

of 17

#3

—

Realtime score

0–100

78.2

—

Time to first word

973ms

—

Reliability

Rolling 30 days

Uptime

100.00%

100.00%

Error rate

0.00%

0.00%

Pricing

billed per second

Rate

per minute

$0.0077

$0.0040

Billing granularity

per 1s

per 1s

1 hour of audio

$0.46

$0.24

100 hours

$46.08

Capabilities

Feature support

Speaker diarization

Yes

Yes

Word-level timestamps

Yes

Yes

Code-switching

Yes

Coverage & deployment

Languages · formats · regions

Languages

47

76

Mode

Batch & Realtime

Batch

Region

United States

United States

Audio formats

general

23.4%
5.4%

legal

3.8%
2.8%

medical

9.5%
5.8%

noisy

21.8%
33.8%

technical

2.8%
3.1%

uk

6.0%
0.0%

us

11.3%
0.0%

Latency p99

1.8s

4.0s

Final drain

181ms

—

Real-time factor

1.01×

—

Flicker

9.6%

—

Cadence

0.8/s

—

$24.12

1,000 hours

$460.80

$241.20

Yes

Auto-detect language

Yes

Yes

Live / streaming

Yes

No

Custom vocabulary

Yes

No

mp3
wav
flac
m4a
ogg
webm
mp3
wav
flac
m4a
ogg
aac

Max file size

2 GB

3 GB

Max duration

No limit

10 h

Compliance

HIPAA
SOC2
GDPR
SOC2
GDPR

Frequently asked questions

Is Nova-3 or Scribe v2 more accurate?

Scribe v2 is more accurate, with a 9.2% word error rate versus 21.0% for Nova-3, on our standardized benchmark.

Which is cheaper, Nova-3 or Scribe v2?

Scribe v2 is cheaper at $0.0040/min versus $0.0077/min for Nova-3.

Should you choose Nova-3 or Scribe v2?

Scribe v2 is the stronger choice overall, leading on accuracy, speed, lower cost, and language coverage.