Cartesia's newest streaming STT — lowest WER of any streaming model, native turn detection, and robust on alphanumerics like phone numbers, emails, and UUIDs. English only for now.
1 languages · not yet benchmarked
This is a realtime-only model. Batch benchmarks are not applicable — realtime performance metrics require a different evaluation methodology.
pay per second
Rate
$0.0065
/min
Per-second billing. Bring your own provider key and pay your provider directly — a 5% routing fee applies (first 100 min/mo free).
Set up BYOKCost estimator
1 hour
$0.39
10 hours
$3.89
100 hours
$38.88
1,000 hours
$388.80
Billed per second of audio processed.
Supported Languages
Supported Formats
Features
Ink-2 is a speech-to-text model from Cartesia. It supports 1 language and runs at $0.006/min — mid-range on price.
Its nearest benchmarked alternative is Chirp 3. Best suited for live captioning and streaming transcription.
