mlsanityData to 7 Oct 2026

ZONOS2 Technical Report

arXiv 2606.24320 ↗First posted 23 Jun 2026cs.SD cs.AIcited by 0

On the map: Mixture-of-experts → Multimodal MoE models

Abstract (arXiv)

We present ZONOS2 8B, our latest TTS model, which achieves state-of-the-art naturalness, prosody, and voice cloning fidelity. We improve upon Zonos-v0.1 across scale, data, and training recipe. We scale the model from 1.6B to 8B total parameters (900M active) with a novel mixture-of-experts (MoE) backbone, improving inference latency and throughput. We expand our training corpus from 200K to over 6M hours using a new data processing pipeline, and we simplify our post-training and conditioning recipes to improve naturalness and voice cloning fidelity. We evaluate ZONOS2 8B on quality, speaker similarity, WER, and ZTTS1-Eval, our novel TTS benchmark, where it performs competitively with state-of-the-art systems while maintaining good streaming latency. We release our model weights and example inference code under an Apache 2.0 license on GitHub and Hugging Face.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Multimodal MoE models, one of 12 directions in Mixture-of-experts. Builds on 28 of its 36 references that are in the map. First posted on arXiv 23 Jun 2026.

Builds on

28 of 36 references.

References · 28

  1. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
  2. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 3,972
  3. 2020GLU Variants Improve Transformercited by 1,724
  4. 2021WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processingcited by 1,609
  5. 2019Root Mean Square Layer Normalizationcited by 1,425
  6. 2019Common Voice: A Massively-Multilingual Speech Corpuscited by 1,358
  7. 2023GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpointscited by 1,311
  8. 2021SoundStream: An End-to-End Neural Audio Codeccited by 756
  9. 2020ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verificationcited by 606
  10. 2023High-Fidelity Audio Compression with Improved RVQGANcited by 423
  11. 2025Muon is Scalable for LLM Trainingcited by 381
  12. 2024Seed-TTS: A Family of High-Quality Versatile Speech Generation Modelscited by 293
16 more
  1. 2020Query-Key Normalization for Transformerscited by 266
  2. 2025Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Freecited by 266
  3. 2024Auxiliary-Loss-Free Load Balancing Strategy for Mixture-of-Expertscited by 191
  4. 2025CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-trainingcited by 142
  5. 2022Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognitioncited by 139
  6. 2020Universal Phone Recognition with a Multilingual Allophone Systemcited by 81
  7. 2024DenseFormer: Enhancing Information Flow in Transformers via Depth Weighted Averagingcited by 48
  8. 2025MiniMax-Speech: Intrinsic Zero-Shot Text-to-Speech with a Learnable Speaker Encodercited by 44
  9. 2020DiDiSpeech: A Large Scale Mandarin Speech Corpuscited by 33
  10. 2024Reshape Dimensions Network for Speaker Recognitioncited by 25
  11. 2024VoxBlink2: A 100K+ Speaker Recognition Corpus and the Open-Set Speaker-Identification Benchmarkcited by 24
  12. 2025Granary: Speech Recognition and Translation Dataset in 25 European Languagescited by 21
  13. 2024FLEURS-R: A Restored Multilingual Speech Corpus for Generation Taskscited by 13
  14. 2025TTSDS2: Resources and Benchmark for Evaluating Human-Quality Text to Speech Systemscited by 7
  15. 2025Measuring Prosody Diversity in Zero-Shot TTS: A New Metric, Benchmark, and Explorationcited by 7
  16. 2025Multi-Sampling-Frequency Naturalness MOS Prediction Using Self-Supervised Learning Model with Sampling-Frequency-Independent Layercited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Tacit-TTS: From Autoregressive Decoding to Masked Prediction for Efficient Transcript-Free Voice Cloning
  2. 2026Adding Voice Cloning to Text-to-Audio-Video Models with a Single Zero-Initialised Layer
  3. 2026CuteTTS: Efficient and High-Quality Speech Synthesis via Autoregressive Modeling of Continuous Latents
  4. 2026Raon-OpenTTS: Open Models and Data for Robust Text-to-Speech ↗
  5. 2023NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers
  6. 2024E2 TTS: Embarrassingly Easy Fully Non-Autoregressive Zero-Shot TTS
  7. 2026PFluxTTS: Hybrid Flow-Matching TTS with Robust Cross-Lingual Voice Cloning and Inference-Time Model Fusion
  8. 2026One Voice, Many Tongues: Cross-Lingual Voice Cloning for Scientific Speech