ElevenLabs, a leader in voice generation technology, is actively demonstrating the capabilities of its new flagship model – Eleven v3 (Alpha) – on June 10-11, 2025. The alpha release of this product occurred on June 8 and has already generated significant buzz within the developer and content creator communities. This launch, as noted by leading tech publications including The Verge and Tom's Guide, marks a significant leap in the quality and capabilities of AI-powered human speech generation. The key improvement in Eleven v3 is its unprecedented emotional expressiveness and intonation control. The AI can generate speech with a wide range of emotions, from joy and enthusiasm to sadness and whispering, making the voice nearly indistinguishable from a live human performance. Furthermore, the model has gained support for over 70 languages, greatly expanding its global applicability. Also introduced were multi-speaker functions for creating realistic dialogues with several unique voices in a single audio track, and support for special audio tags (e.g., [laughs] or [pause]) for fine-grained control over the generation process. Despite its alpha status, initial feedback from the community has been enthusiastic. One Reddit user enthusiastically commented, "When it gets it right, the generations are breathtaking!" The release of Eleven v3 sets a new quality standard in the voice synthesis industry and opens new horizons for audiobook creators, podcasters, game developers, and accessibility technology specialists.
ElevenLabs Demonstrates Breakthrough Eleven v3 (Alpha) Voice Synthesis Model
