# ZONOS3-Mini Technical Report

**Authors:** Gabriel Clark, Sofian Mejjoute, George Close, Beren Millidge  
**Lab:** Hugging Face · **Published:** 12 Nov 2026  
**Links:** [arXiv](https://arxiv.org/abs/2611.16525) · [PDF](https://arxiv.org/pdf/2611.16525) · [This page](https://omen-bit.github.io/ai-paper-daily/paper/2611.16525.html)

## Key numbers

- **6.71%**: ZONOS3-Mini reaches a 6.71% macro-average WER over 55 language sets
- **886M**: To achieve this we propose a latent MoE backbone with only 886M parameters

## Abstract

ZONOS3-Mini is an ultra-compact, ultra-fast, zero-shot voice-cloning TTS model. The model can run in real time on a laptop or smartphone. For its latency and size, it achieves state-of-the-art accuracy and cloning fidelity. To achieve this we propose a latent MoE backbone with only 886M parameters (100M active per token) and a new audio codec, Zyphra Audio Codec 1 (ZAC-1), which represents 44.1 kHz audio at a 15 Hz frame rate. On the new ZTTS2-Eval benchmark, ZONOS3-Mini reaches a 6.71% macro-average WER over 55 language sets (35 clean and 20 in the wild), which is competitive with larger and slower state-of-the-art models. We release the model weights, inference, and training code.

---

Notes written by AI from the paper, numbers copied from the abstract. Check the paper before you rely on them. From AI Papers Daily.