ElevenLabs has launched Eleven v4 and Eleven v4 Turbo, two new speech models designed to provide more expressive voice generation, faster cloning, and lower latency for real-time voice agents. The models support more than 90 languages, up from 70 in the previous generation.
The company announced Eleven v4 on September 28, following the release of v3 last year and an earlier preview at an event in Warsaw. ElevenLabs said the new architecture improves control over speech and allows users to clone a voice from as little as 10 seconds of audio.
More Control Over Expression and Voice Identity
Eleven v4 is designed to interpret tone, pacing, emotion, character, and context while preserving a speaker’s identity across longer passages. The model can adjust how text is spoken based on surrounding context and supports improved multi-speaker dialogue.
ElevenLabs is also expanding the inline expression tags introduced with v3. Users can combine multiple tags and have the model follow them in sequence, providing more detailed control over how generated speech is delivered.
Language support has increased to more than 90 languages, compared with more than 70 for v3. ElevenLabs said some of the largest quality improvements were seen in Japanese, Brazilian Portuguese, Mandarin, and Cantonese.
v4 Turbo Targets Real-Time Voice Agents
Eleven v4 Turbo is designed for real-time speech generation and has median inference latency of about 100 milliseconds, according to ElevenLabs documentation. The company positions it for uses including support agents, AI assistants, and interactive characters.
The new generation can begin generating audio while the underlying large language model is still producing its response. ElevenLabs also said the models can handle situations such as confrontations, escalations, and holds differently to support more natural conversations.
The launch comes as ElevenLabs expands its enterprise business, with more than 55% of its business now coming from large companies, according to TechCrunch. The company faces competition in speech generation from startups including Cartesia, Deepgram, Fish Audio, Boson, and WellSaid Labs, along with Google and OpenAI.
ElevenLabs raised $500 million from Sequoia earlier this year at an $11 billion valuation, while reports have pointed to another possible round at a $22 billion valuation. Its annualized revenue run rate has increased from about $330 million at the start of the year to more than $600 million, while headcount has passed 800 as it hires across markets including India, Europe, and Brazil.
Co-founder and CEO Mati Staniszewski recently said ElevenLabs is aiming for an IPO “in the next years,” but did not provide a timeline.
Featured image credits: Wikimedia Commons
For more stories like it, click the +Follow button at the top of this page to follow us.