ElevenLabs launches v4 speech models in over 90 languages

London-based AI voice startup ElevenLabs launched its v4 and v4 Turbo speech models on September 28. The new models offer improved expression controls, lower latency for voice agents, and support for more than 90 languages.

The company stated that the models can clone a voice from just 10 seconds of audio, maintain a more consistent voice identity over longer passages, and begin generating audio while a large language model is still producing text.

ElevenLabs indicated that this release expands language coverage from 70 languages in v3 to over 90. Significant quality gains were reported for Japanese, Brazilian Portuguese, Mandarin, and Cantonese.

The company, which also offers an enterprise AI voice platform, noted that large corporations account for over 55% of its business. In February, ElevenLabs announced it had raised $500 million at an $11 billion valuation.

The launch enters a competitive landscape with rivals such as Cartesia and Deepgram also targeting the voice agent and enterprise text-to-speech markets.