Cartesia
Ultra-low-latency voice models built on state space models.
Full Cartesia record →Head to head
Cartesia (Cartesia) and Deepgram (Deepgram) both sit in Audio. Cartesia is the cheaper entry point at Free tier · from $5/mo. Deepgram shipped more tracked changes in the last 30 days (5 vs 0). Every value below comes from the latest crawl of the vendors' own pages.
| Attribute | Cartesia Cartesia | Deepgram Deepgram |
|---|---|---|
| Segment | Audio | Audio |
| Version | Sonic 3 | Nova-3 |
| Entry cost | Free tier · from $5/mo | Free tier · from $4,000/min |
| Pricing tiers | — | Free free · Pay As You Go free · Growth $4000 · Enterprise free |
| Model stack | Sonic 3 | Aura-1 · Aura-2 · Flux · Flux English · Flux Multilingual · Flux TTS · Nova-3 · Nova-3 Monolingual · Nova-3 Multilingual · Whisper · Whisper Cloud · Whisper Large |
| Context window | — | n/a |
| Public API | ||
| Routes models | ||
| Changes / 30d | 0 | 5 |
| Origin | Global | Global |
| Capabilities | Sub-100ms TTS, Voice cloning, Realtime API | Real-time streaming, Keyterm prompting, Multilingual, Voice agent API, Text to speech, Speech-to-Text, Text-to-Speech, Voice Agent API |
Ultra-low-latency voice models built on state space models.
Full Cartesia record →Low-latency streaming transcription and voice agent infrastructure.
Full Deepgram record →