Eleven v4 and v4 TurbovsCartesia Sonic

Eleven v4 Turbo and Cartesia are the two text-to-speech options voice-agent builders now weigh for live conversation. Eleven v4 arrived on September 28, 2026 with a quality model for produced audio and a Turbo model at about 100 ms median latency. Cartesia built its reputation on ultra-low latency and streaming from the start, with a smaller catalogue of voices.

At a glance

Eleven v4 and v4 TurboCartesia Sonic
CategoryVoiceVoice
FunctionText-to-speech modelVoice gen
PricingFreemiumFreemium
Starting priceIncluded in ElevenLabs plans; free tier with limited characters, paid from $5/moFree 10k chars/mo, Pro from $49/mo
Editor rating—4.6 / 5
ReleasedSeptember 2026May 2024

Eleven v4 and v4 Turbo

Pros

  • Directions in the script replace fiddly per-line settings
  • Two models cover both produced audio and real-time agents
  • Available to existing subscribers at no new price

Cons

  • Latency figures are the vendor's own
  • Character-based pricing still adds up for long content
  • Voice cloning rules and consent checks apply as before

Cartesia Sonic

Pros

  • Lowest-latency realistic voice AI
  • Quality matches ElevenLabs
  • Streaming API is exceptionally good

Cons

  • Pricier than alternatives for batch use
  • Voice library smaller than ElevenLabs

Which one should you pick?

Pick Eleven v4 and v4 Turbo if

You need one vendor for both narration and live agents, want 90+ languages and inline direction tags, or already run ElevenLabs.

See Eleven v4 and v4 Turbo details →

Pick Cartesia Sonic if

Latency is the only metric that matters, you want fine control over streaming and interruption handling, or price per character at scale favours it for your volume.

See Cartesia Sonic details →

Bottom line

Eleven v4 for breadth and produced quality with a Turbo tier that now competes on speed. Cartesia for builders who optimise every millisecond. Test both on your own agent before committing; latency figures are the vendors' own.