LongCat-AudioDiT
Meituan · China
Meituan diffusion text-to-speech model
- Release month
- Field
- Voice
- Availability
- open-weight
- Parameters
- 1B, 3.5B
- License
- mit · Commercial OK
- Size
- 1B · 3.5B
Overview
- Registered on Hugging Face March 30, 2026 · Open weights · 1B and 3.5B
- Text-to-speech · Zero-shot voice cloning · Generates directly in waveform latent space
View official announcement ↗
← All models