2 models compared on the same benchmark — accuracy, latency, price, language coverage and capabilities. Soniox STT (Async) leads on raw accuracy; Soniox STT (Async) is fastest; Nova-2 covers the most languages; Soniox STT (Async) is cheapest.
Verdict · who wins what
Most accurate
Soniox STT (Async)
8.7% WER
Soniox
Fastest
Soniox STT (Async)
0.1× realtime
Soniox
Most languages
Nova-2
33 languages
Deepgram
Cheapest
Soniox STT (Async)
$0.0017/min
Soniox
Comparing
2 models · same benchmark
1
Nova-2
Deepgram
Deepgram's Nova-2 speech recognition
2
Soniox STT (Async)
Soniox
Soniox async batch transcription — one multilingual model across 60+ languages with speaker diarization, per-word timestamps, and automatic language identification, at a low flat token-based rate.
30-day benchmark average
overall WER vs price
Nova-2
Soniox STT (Async)
field · 33 models
lower-left is better
WER · lower is better
WER · English
p50 → p99
Streaming benchmark averages
Rolling 30 days
billed per second
Feature support
Languages · formats · regions
general
legal
medical
noisy
technical
uk
us
Latency p99
3.7s
6.2s
Final drain
192ms
—
Real-time factor
1.02×
—
Flicker
10.5%
—
Cadence
0.8/s
—
$10.08
1,000 hours
$349.20
$100.80
No
Auto-detect language
Yes
Yes
Live / streaming
Yes
No
Custom vocabulary
Yes
No
Max file size
2 GB
No limit
Max duration
No limit
5 h
Compliance
—
Is Nova-2 or Soniox STT (Async) more accurate?
Soniox STT (Async) is more accurate, with a 8.7% word error rate versus 16.5% for Nova-2, on our standardized benchmark.
Which is cheaper, Nova-2 or Soniox STT (Async)?
Soniox STT (Async) is cheaper at $0.0017/min versus $0.0058/min for Nova-2.
Should you choose Nova-2 or Soniox STT (Async)?
Choose Nova-2 for language coverage; choose Soniox STT (Async) for accuracy, speed, and lower cost.
