ElevenLabs Launches v4 Speech Models With More Expression Control and 90+ Languages

ElevenLabs has launched two new speech models, ElevenLabs v4 and v4 Turbo, adding more control over expression, lower latency for voice agents and support for more than 90 languages.

The company released its v3 model last year and previewed the next generation at an event in Warsaw earlier this year. For v4, ElevenLabs is using a new architecture designed to improve voice control and make cloning faster.

The company says users can clone a voice with as little as 10 seconds of audio. The model is also designed to preserve voice identity more consistently across longer passages.

More Control Over Voice Expression

ElevenLabs v4 can use the context of written text to adjust how speech is delivered. The company has also expanded the inline expression tags introduced with v3, allowing users to combine multiple tags and have the model follow them in sequence.

Language support has increased from 70 languages in the previous generation to more than 90. ElevenLabs said some of the largest quality improvements are in Japanese, Brazilian Portuguese, Mandarin and Cantonese.

Lower Latency for AI Voice Agents

ElevenLabs is also positioning v4 for conversational voice agents. The model has lower latency and can begin producing audio while the underlying large language model is still generating a response, making conversations feel more immediate.

The company said the system can also adjust its behaviour during situations such as confrontations, escalations and holds to help voice agents handle customer interactions more effectively.

Enterprise calling has become an increasingly important part of ElevenLabs’ business, with more than 55% of its business now coming from large companies.

Competition in AI Speech Models Intensifies

Competition in speech generation continues to grow as startups including Cartesia, Deepgram, Fish Audio, Boson and WellSaid Labs develop expressive voice models. Google and OpenAI have also continued improving their voice technologies.

ElevenLabs raised $500 million earlier this year in a Sequoia-led round that valued the company at $11 billion. Its annualised revenue run rate has grown from about $330 million at the start of the year to more than $600 million, while its workforce has expanded to more than 800 people.

The company has been hiring across markets including India, Europe and Brazil. Co-founder and CEO Mati Staniszewski has said ElevenLabs is aiming for an IPO “in the next few years,” without committing to a specific timeline.