A new, open source text-to-speech model called Dia has arrived to challenge ElevenLabs, OpenAI and more

Nari Labs' Dia, a cutting-edge text-to-speech model with over 1.6 billion parameters, surpasses existing AI tools like ElevenLabs and Google's NotebookLM in natural dialogue, emotional expression, and nonverbal cues. Built using Google's TPU chips for efficiency, Dia supports customizable features such as voice tags and audio conditioning, making it highly versatile. Nari Labs plans to release a consumer-friendly version soon, with the model open source under Apache 2.0, while also exploring applications in creative and assistive technologies.