ElevenLabs releases v4 speech model with faster voice agents

elevenlabs released eleven v4, a speech model that follows direction cues more accurately and keeps voices consistent across long productions. the model supports over 90 languages, improves voice cloning, and handles up to 10,000 characters per request. eleven v4 turbo, a variant for real-time voice agents, produces speech in 150 milliseconds, outperforming competing models like cartesia sonic 3.6 and openai's gpt-4o mini tts. the update affects voice agent developers and users seeking more expressive and consistent ai voices, with potential applications in dubbing, customer service calls, and game characters.