Compare realtime APIs

Pick up to four APIs from any category and see them side by side.

Cartesia Sonic
Cartesia
Gradium TTS
Gradium
Rime TTS (Coda, Mist v3)
Rime Labs
CategoryText-to-speechText-to-speechText-to-speech
StatusGABetaGA
Est. per minute$0.034 - 0.059$0.032 - 0.052$0.027 - 0.045
How that was worked out900 chars/min at 1 credit/char. Low = Scale plan effective ~$37.4/1M; high = Pro overage $65/1M.900 chars/min; L plan ~$35.9/1M vs XS ~$57.8/1M900 chars/min; Mist v3 vs Coda at Starter rates
Pricing modelsubscriptionsubscriptionper-character
Free tier20K credits/month, overages blocked; commercial use not listed for Free (Pro and up: yes).45K credits/month, non-commercialPage conflicts: '~800 minutes free (about 800k characters)' vs FAQ '3,000 free minutes'
Connects byWebSocket, HTTP chunked, SSEWebSocketWebSocket, HTTP chunked
Audio inText (no SSML required; break tags supported)TextText (custom pronunciation via spell() and dictionaries)
Audio outRaw container; pcm_f32le, pcm_s16le, pcm_mulaw, pcm_alaw at 8000/16000/22050/24000/44100/48000 Hz on the WebSocketNot verifiedpcm, mp3, mulaw (audioFormat); sample rates per model reference
Languages44 (Sonic 3.6)English, French, German, Spanish, Portuguese (AWS Marketplace listing)Coda: 8; Mist v3: 4
Latency (vendor claim)Third-party sources quote Cartesia's claims of ~90 ms time-to-first-audio for Sonic 3 and ~40 ms for Sonic Turbo; Sonic 3.6 described as sub-90 ms. Independent benchmarks reported 128-166 ms medians including network for Sonic 3/3.5.Vendor: ~200 ms first audio; vendor-reported Coval P50 TTFA 158 ms.Vendor: sub-200 ms end-to-end via cloud API; Mist v3 typical TTFB well below 100 ms; Coda sub-100 ms model latency on GPU engine (self-hosted) plus 25-50 ms network from most of the US.
Key limits
  • TTS concurrency: Free 2, Pro 3, Startup 5, Scale 15
  • max_buffer_delay_ms default 3000 (0-5000)
  • voice and output_format must stay constant within a context
  • Overages must be enabled; Free stops at the credit cap
  • TTS concurrency: Free 2, XS/S 5, M 10, L 15
  • Starter: 20 concurrent TTS generations
  • Rime keeps only one active contextId at a time per socket
High-severity warnings
  • None
  • Free tier is non-commercial
  • Arcana is gone from the cloud
ComplianceNot verified in this pass (enterprise custom terms available).Not verified.Enterprise: BAA and SOC 2 reports; optional request-data retention controls (disabled by default).
Self-hostableNoNoYes
Last checked2026-10-102026-10-102026-10-10
Key numbers and features
$/1M chars$50$57.8$50
Free tierYesYesYes
Free credit $---
Free tier commercialNoNo-
Voices---
CloningYes--
Instant cloneYes--
Text stream inYesYesYes
TimestampsYes-Yes
EmotionYes--
SSML---
8 kHz phoneYes-Yes
Latency ms90200200
Languages4458
Max session min---
Concurrency3520
WebRTCNoNoNo
WebSocketYesYesYes
gRPCNoNoNo
HIPAA--Yes
SOC 2--Yes
EU data---
Self-hostNoNoYes
Open weightsNoNoNo
High warnings011