ElevenLabs Eleven v3 Reaches General Availability With 74 Languages and Audio Tags

Audio Tags let you write emotion and delivery cues inline in your script -- [angry], [whisper] or [laughing] drop straight into the text and the model handles the performance shift without a separate take or manual clip editing. The model first launched in alpha in June 2025; GA brings it to all paid creator plans. For YouTube creators and course producers targeting non-English markets, 74-language support means a single voice identity can follow your content into languages where you previously had to hire local voice actors or accept a noticeably different persona. Professional Voice Cloning, available on creator-tier accounts and above, changes the cost equation significantly: a creator who previously needed a recording studio and a licensing agreement can now build a voice asset from their own existing audio. The practical test for an upgrade is whether you need Audio Tags and 74-language output together regularly -- that combination is where the higher plan starts earning its price difference over the Starter tier. See ElevenLabs' pricing page for current plan options.