Key facts
- ElevenLabs launched new v4 and v4 Turbo speech models.
- The new models offer more expression control and lower latency.
- Support for over 90 languages is included, up from 70 previously.
- Voice cloning requires only 10 seconds of audio with the v4 models.
- The company's annualized revenue run rate has increased to over $600 million.
- ElevenLabs has over 800 employees.
ElevenLabs has introduced its latest speech models, v4 and v4 Turbo, enhancing capabilities such as expression control, latency for voice agents, and language support. The new architecture allows for voice cloning with as little as 10 seconds of audio and better handling of longer text segments by maintaining context for more natural expression changes. The number of supported languages has increased from 70 to over 90, with notable quality improvements observed in Japanese, Brazilian Portuguese, Mandarin, and Cantonese. The company highlighted the models' suitability for voice agents due to reduced latency, enabling more fluid conversations. ElevenLabs has experienced rapid growth in its enterprise calling business, with over 55% of its revenue coming from large corporations. The company has also seen its annualized revenue run rate climb to over $600 million from approximately $330 million at the start of the year, and its headcount has surpassed 800 employees. Earlier this year, ElevenLabs secured $500 million in funding led by Sequoia, valuing the company at $11 billion, with reports of a potential follow-up round at a $22 billion valuation. CEO Mati Staniszewski indicated a potential IPO within the next few years.