light estimateLast updated 2026-08-20

Text-to-speech API with voice cloning

Default pickElevenLabs Multilingual v2 / Eleven v3

For voice cloning, ElevenLabs is the most established option: the largest voice library, the most mature instant and professional cloning workflows, and streaming support. Cartesia offers instant and professional cloning on its paid plans and MiniMax supports voice replication, while Google Gemini and OpenAI's gpt-4o-mini-tts do not clone voices. Pricing and quality differ sharply by tier, so check each provider row. A light estimate from provider documentation, August 2026.

DefaultElevenLabs Multilingual v2 / Eleven v3stemklonenElevenLabs Multilingual v2 / Eleven v3
Provider offerings compared on Price, ELO, TTFA, Langs and capabilities
OfferingPrice ($/1M chars)ELOTTFALangsCapabilities
ElevenLabs Multilingual v2 / Eleven v3ElevenLabs1001179653 ms29streamingcloning
Cartesia Sonic 3 / Sonic 3.5Cartesia39*1203271 ms42streamingcloning
Google Gemini 3.1 Flash TTSGoogle12*1212streaming
OpenAI gpt-4o-mini-ttsOpenAI15*775 ms50streaming
MiniMax Speech 2.5 TurboMiniMax78streamingcloning
Deepgram Aura-2Deepgram30293 ms7streaming
SpeechifyAI Simba 3.2Speechify10*1240398 ms30streamingcloning

* token-/credit-priced — the headline understates real per-unit cost, so it is excluded from the cheapest ranking.