Cartesia matches Gemini 3.8 TTS on quality and generates speech about twice as fast.

Artificial Analysis times generation for every model on its board. As of October 2026, Sonic 3.6 generates about 106 characters of speech per second, Gemini 3.8 Flash-Lite TTS about 55, and Gemini 3.8 Flash TTS about 47. On the same board's blind listening test, Sonic 3.6 and Gemini 3.8 Flash TTS score within 3 points of each other. For a voice agent, generating well ahead of playback is what keeps a long reply from stalling mid-sentence.

Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, which Google released in September 2026, and Gemini 3.1 Flash TTS, the model they replace. Those are the names Artificial Analysis lists them under. Google's model IDs are gemini-3.8-flash-tts, gemini-3.8-flash-lite-tts, and gemini-3.1-flash-tts-preview.

Google sells an older set of Cloud Text-to-Speech voices alongside Gemini TTS: Chirp 3: HD, Studio, Journey, Neural2, WaveNet, and Standard. Artificial Analysis rates every one of them below the Gemini TTS models on this page.

It comes from Artificial Analysis, an independent benchmark. Listeners hear two clips of the same line without being told which model produced either, and pick the one they prefer. The Provider Voice Arena lets every model speak in its own voices.

Yes. An instant clone takes 10 seconds of audio and a professional clone takes 30 minutes. Both are on the public API, rather than reserved for an enterprise agreement.

Yes. Cartesia deploys on-prem and air-gapped, which teams in healthcare, finance, and government use to keep audio inside their own network, backed by a 99.9% uptime SLA.

One to two weeks for most teams. Cartesia is integrated into the major voice agent platforms and every model is on the public API, so there is no platform to move onto and nothing to unpick later.