# Audio8-TTS-Realistic-Air

> This release supports 23 languages, adding 12 languages compared with Preview. It uses the same Transformers inference interface as Preview. This repository includes the complete model checkpoint, its 44.1 kHz neural audio codec, tokenizer, processor, and custom Transformers code.

- **Organisation:** Edge0
- **Category:** Text-to-Speech
- **Parameters:** 601M
- **Licence:** Apache 2.0 (Commercial use OK)
- **Released:** 2026-10-14
- **Hugging Face:** 25.6K downloads (30 days), 480 likes

## Estimated memory

| Quant | ≈ GB |
|---|---|
| Q4_K_M | 0.91 |
| Q8_0 | 1.21 |
| F16 | 1.84 |

## Runs with

- [Transformers](https://github.com/huggingface/transformers)
- [MLX](https://github.com/ml-explore/mlx-lm) (community)

## About

This release supports 23 languages, adding 12 languages compared with Preview. It uses the same Transformers inference interface as Preview. This repository includes the complete model checkpoint, its 44.1 kHz neural audio codec, tokenizer, processor, and custom Transformers code.

Try speech generation and voice cloning in the interactive Playground. Hosted on Hugging Face ZeroGPU; availability depends on the queue and your daily GPU allowance.

• Improved realism: more natural prosody and expressive delivery, building on Preview.
• Broader multilingual coverage: 23 languages, including 12 additions to Preview's 11 supported languages.
• Compact model: 601 million main-model parameters, with the audio codec bundled separately.
• Flexible generation: zero-shot voice cloning from a reference recording and transcript, or speech generation without a reference.
• GPU serving: an SGLang Omni adapter with paged attention, dynamic batching, and an OpenAI-compatible speech API.

Note: Language quality varies; see the per-language evaluation below. Benchmark coverage includes some additional languages and is distinct from the supported-language list.

Audio8 TTS uses a DualAR architecture inspired by Fish Audio S2 Pro. The slow AR transformer predicts one semantic token for each audio frame. The fast AR transformer predicts the frame's codec codebooks, conditioned on the slow hidden state and preceding codebooks.

## Alternatives

- [k2-fsa/OmniVoice](https://selfhost.directory/models/k2-fsa/omnivoice) (613M, Unknown licence, ≈1.87 GB): OmniVoice is a massively multilingual zero-shot text-to-speech (TTS) model supporting over 600 languages. Built on a novel diffusion language model-style architecture, it delivers high-quality speech with superior inference speed, supporting voice cloning and voice design.
- [Edge0/Audio8-TTS-Realistic-0.6b](https://selfhost.directory/models/edge0/audio8-tts-realistic-0.6b) (601M, Apache 2.0, ≈1.84 GB): This release supports 23 languages, adding 12 languages compared with Preview. It uses the same Transformers inference interface as Preview. This repository includes the complete model checkpoint, its 44.1 kHz neural audio codec, tokenizer, processor, and custom Transformers code.
- [ResembleAI/chatterbox](https://selfhost.directory/models/resembleai/chatterbox) (500M, MIT, ≈1.62 GB): Chatterbox Multilingual V3 is the latest general-purpose multilingual TTS model in the Chatterbox family. It keeps the same 0.5B model size while improving speaker similarity, reducing hallucinations, and producing more natural, conversational speech across languages.
- [FunAudioLLM/Fun-CosyVoice3-0.5B-2512](https://selfhost.directory/models/funaudiollm/fun-cosyvoice3-0.5b-2512) (500M, Apache 2.0, ≈1.62 GB): Fun-CosyVoice 3.0 is an advanced text-to-speech (TTS) system based on large language models (LLM), surpassing its predecessor (CosyVoice 2.0) in content consistency, speaker similarity, and prosody naturalness. It is designed for zero-shot multilingual speech synthesis in the wild.
- [nineninesix/gepard-1.0](https://selfhost.directory/models/nineninesix/gepard-1.0) (556M, Apache 2.0, ≈1.74 GB): Gepard is a text-to-speech model built for real-time conversation. It starts speaking the moment text begins arriving, generating audio piece by piece instead of waiting for a full sentence — so it feels like a live voice, not a recording.
- [Qwen/Qwen3-TTS-12Hz-0.6B-Base](https://selfhost.directory/models/qwen/qwen3-tts-12hz-0.6b-base) (915M, Apache 2.0, ≈2.54 GB): Qwen3-TTS is a family of advanced multilingual, controllable, robust, and streaming text-to-speech models. Trained on over 5 million hours of speech data spanning 10 languages, Qwen3-TTS supports state-of-the-art 3-second voice cloning and description-based control.

More: https://selfhost.directory/models/edge0/audio8-tts-realistic-air/alternatives

## Runs well on

- [GeForce RTX 4060](https://selfhost.directory/models/edge0/audio8-tts-realistic-air/on/rtx-4060): F16
- [Apple M4 (16 GB)](https://selfhost.directory/models/edge0/audio8-tts-realistic-air/on/apple-m4-16gb): F16
- [Radeon RX 6700 XT](https://selfhost.directory/models/edge0/audio8-tts-realistic-air/on/rx-6700-xt): F16
- [GeForce RTX 3060 12GB](https://selfhost.directory/models/edge0/audio8-tts-realistic-air/on/rtx-3060): F16
- [Arc B580](https://selfhost.directory/models/edge0/audio8-tts-realistic-air/on/arc-b580): F16
- [Tesla T4](https://selfhost.directory/models/edge0/audio8-tts-realistic-air/on/tesla-t4): F16

## Similar size

- [k2-fsa/OmniVoice](https://selfhost.directory/models/k2-fsa/omnivoice) (613M)
- [ResembleAI/chatterbox](https://selfhost.directory/models/resembleai/chatterbox) (500M)
- [FunAudioLLM/Fun-CosyVoice3-0.5B-2512](https://selfhost.directory/models/funaudiollm/fun-cosyvoice3-0.5b-2512) (500M)
- [coqui/XTTS-v2](https://selfhost.directory/models/coqui/xtts-v2) (467M)
- [Edge0/Audio8-TTS-Realistic-0.6b](https://selfhost.directory/models/edge0/audio8-tts-realistic-0.6b) (601M)
- [openbmb/VoxCPM-0.5B](https://selfhost.directory/models/openbmb/voxcpm-0.5b) (500M)

## FAQ

**How much VRAM does Audio8-TTS-Realistic-Air need?**
About 1.8 GB at FP16 / BF16.

**Can Audio8-TTS-Realistic-Air run on the RTX 4060?**
Yes. It fits at FP16 / BF16 (about 1.8 GB). It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, RTX 3060 12GB.

**Can I use Audio8-TTS-Realistic-Air commercially?**
Yes. It is released under Apache 2.0. Permissive: use, modify and sell freely; keep the notice and licence text.

**How do I run Audio8-TTS-Realistic-Air locally?**
It runs with Transformers. Each has copy-paste examples in its docs.

## Links

- huggingface: https://huggingface.co/Edge0/Audio8-TTS-Realistic-Air
- github: https://github.com/Edge0-AI/Audio8_TTS
- paper: https://arxiv.org/abs/2610.14776
- demo: https://huggingface.co/spaces/Edge0/Audio8-TTS-Realistic-Air

---
Source: https://selfhost.directory/models/edge0/audio8-tts-realistic-air