ElevenLabs Pitches Eleven v4 at Dubbing and Games With 10-Second Voice Clones

ElevenLabs releases Eleven v4 and Eleven v4 Turbo, text-to-speech models it says are ranked first by benchmarking firm Artificial Analysis and preferred by about 75% of listeners in blind head-to-head tests, with more than 90 languages and instant voice clones built from 10 seconds of audio. The company lists audiobooks, character performances, voiceovers, dubbing, localisation and games among the uses, and puts v4 Turbo's median time to first speech at about 150 milliseconds. Both are available now in its agents and creative products and through its API; the announcement names no price and no customer.