ElevenLabs v4 Model Brings More Expression Control and 90 Languages
ElevenLabs launched two new speech models on Monday, called the ElevenLabs v4 model and v4 Turbo. The new models offer more expression control, lower latency for voice agents, and support for more than 90 languages.
The company released its v3 model last year and teased the new model at an event in Warsaw earlier this year. For the v4 generation, ElevenLabs is adopting a new architecture that allows for better control and faster cloning. The company said that with v4, users will be able to clone a voice with just 10 seconds of audio.
Faster Voice Cloning With Just 10 Seconds of Audio
One of the most striking improvements in the ElevenLabs v4 model is how quickly it can clone a voice. With only 10 seconds of audio, users can create a usable voice clone. That is a significant jump from earlier versions, which typically needed more recorded material to produce a convincing result.
This faster cloning process matters for creators who want to move quickly. Podcasters, video producers, and developers building voice agents can now set up a custom voice in far less time. The new architecture behind the v4 generation is what makes this speed possible.
Better Voice Identity Over Longer Text
On the creative side, the ElevenLabs v4 model handles voice identity better over longer chunks of text. It also keeps the context of the text in mind while reading it aloud to change expressions. That means the voice does not just read words flatly. It adjusts tone and delivery based on what the text is actually saying.
ElevenLabs introduced inline tags to define expression with v3 and is expanding those tags in v4. Users can now stack multiple tags and have the model follow the sequence. This gives creators finer control over how a line should sound, whether it needs to be warm, urgent, calm, or excited.
Language Support Grows to 90 Languages
The previous version supported 70 languages, and ElevenLabs has worked to get that number up to 90 languages with the new version. The startup said it observed the biggest quality jump in Japanese, Brazilian Portuguese, Mandarin, and Cantonese.
That broader language coverage opens the door to more markets and more use cases. A business that serves customers in several regions can now rely on one model for more of its voice needs. The quality improvements in those four languages are especially notable because they are spoken by large audiences worldwide.
Built for Voice Agents and Real Conversations
ElevenLabs has scaled its enterprise calling business rapidly over the last year, with more than 55% of its business coming from large companies. The company said the new model is suited for voice agents, as the new version has lower latency to allow for more fluid conversation.
Lower latency means less awkward silence during a call. Plus, the ElevenLabs v4 model can start generating audio as soon as the LLM behind it starts generating answers. What’s more, the model can also handle confrontations, escalations, and holds differently for better issue resolution.
That last point matters for customer service. A voice agent that can recognize when a caller is upset and adjust its approach is far more useful than one that reads from a fixed script. The v4 model is designed to manage those moments more effectively.
Competition in Speech Models Heats Up
Competition in speech models has ramped up as startups like Cartesia, Deepgram, Fish Audio, Boson, and WellSaid Labs have created expressive speech models. Big companies like Google and OpenAI have also improved their voice models.
This crowded field pushes every player to move faster. For ElevenLabs, the v4 release is a way to stay ahead on expression control, language coverage, and latency at the same time. Each of those areas matters to different kinds of users, from solo creators to large enterprises.
Funding, Revenue, and Growth at ElevenLabs
ElevenLabs raised $500 million earlier this year from Sequoia in a round led by Sequoia that valued the company at $11 billion. There are already rumors of a follow-up fundraising round that would value the company at $22 billion.
ElevenLabs annualized revenue run rate has climbed from roughly $330 million at the start of the year to over $600 million. The company has aggressively hired personnel in different markets such as India, Europe, and Brazil, as its headcount has reached over 800.
In a recent interview, the company’s co-founder and CEO Mati Staniszewski said that the company is aiming for an IPO “in the next years,” but didn’t commit to a timeline. That ambition suggests the company sees long-term growth ahead as voice AI becomes more central to how businesses operate.
What the ElevenLabs v4 Model Means for Users
With the ElevenLabs v4 model, creators and businesses get faster voice cloning, broader language coverage, and more precise expression control. For voice agent developers, the lower latency and ability to handle complex conversations could make AI-powered calls feel more natural and useful.
The race for better speech models is clearly accelerating. This release shows how much ground can be covered in a single generation, from cloning speed to language reach to how well a voice handles real, messy conversations.
