ElevenLabs Unveils Eleven v3: A Leap in AI Text-to-Speech Technology

ElevenLabs has launched Eleven v3, an advanced AI text-to-speech model that promises to set a new standard in the industry, according to ElevenLabs. This alpha version introduces groundbreaking features such as advanced audio tags, dialogue mode, and support for more than 70 languages, aiming to deliver highly expressive and emotionally nuanced speech outputs.

Innovative Features of Eleven v3

Eleven v3 offers a suite of features designed to enhance the realism and versatility of AI-generated speech. The model includes advanced audio tags, enabling precise control over tone, emotion, and non-verbal cues such as sighs and whispers. Its dialogue mode supports multi-speaker interactions with natural pacing and interruptions, providing a more authentic conversational experience. Additionally, the model covers over 70 languages, catering to a diverse global audience.

Addressing Expressiveness in AI Speech

Since the release of its Multilingual v2, ElevenLabs has observed significant adoption of voice AI in various sectors like film, game development, and education. However, the challenge of achieving expressive and dynamic speech remained. Eleven v3 is engineered to overcome these limitations, offering voices that can express emotions like laughter and surprise, enhancing the lifelikeness of AI interactions.

Practical Applications and Future Development

The launch of Eleven v3 opens new possibilities for creators working on projects such as videos, audiobooks, and media tools, allowing for a higher level of expressiveness. While the current version requires more sophisticated prompt engineering than its predecessors, it promises breathtaking results. A real-time version of v3 is currently in development, with v2.5 Turbo or Flash recommended for real-time applications for the time being.

Availability and Pricing

Eleven v3 is now available on the ElevenLabs website and via API. During its initial release phase, users can enjoy an 80% discount on the UI self-serve plans, making it significantly more affordable compared to previous versions. This promotional pricing is set to last until the end of June 2025.

Conclusion

With its pioneering features and enhanced expressiveness, Eleven v3 positions itself as a transformative tool in the realm of AI text-to-speech technology. As developers and creators integrate these capabilities into their projects, the model promises to redefine the boundaries of what is possible with AI-driven audio content.