ElevenLabs Multilingual v2

ElevenLabs Multilingual v2, released in August 2023, was ElevenLabs' flagship text-to-speech model for lifelike, emotionally rich speech synthesis across 29 languages. Built for voiceovers, audiobooks, and long-form content, it became known for stable output on extended generations and consistent speaker characteristics across multilingual output. It supports expressive voice cloning, letting creators reproduce a specific voice's tone and delivery from a short sample.

The model is closed-weights and proprietary, accessible only through ElevenLabs' API and platform, and it has since been joined by newer options: Flash v2.5 for lower-latency streaming and the v3 flagship with 74-language coverage. Multilingual v2 remains the quality-first choice in the lineup for projects where natural prosody matters more than speed. Third-party pricing aggregators list it around $100 to $234 per million characters depending on the reseller, versus ElevenLabs' own published rates.