{"id":"edge0-audio8-tts-realistic-0-6b","slug":"edge0-audio8-tts-realistic-0-6b","name":"Audio8-TTS-Realistic-0.6B","orgId":"edge0","family":"edge0-audio8-tts-realistic-0-6b","releaseDate":null,"availability":"announced","openWeights":true,"license":null,"paramsTotal":601159424,"paramsSource":"tracked","contextWindow":null,"knowledgeCutoff":null,"modelCardUrl":"https://x.com/SamuelZengML/status/2115881491645554870","hfRepo":null,"modalities":["audio"],"aliases":["Audio8-TTS-Realistic-0.6B","audio8-tts-realistic","audio8 tts realistic","audio8-tts-realistic-0.6b","Edge0/Audio8-TTS-Realistic-0.6b","edge0/audio8-tts-realistic-0.6b","edge0 audio8 tts realistic"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-10-29","announced":"2026-10-29","blurb":"Edge0's open-weights text-to-speech model with zero-shot voice cloning from a short reference clip and its transcript, or speech without a reference. 601M parameters plus a bundled 44.1 kHz codec, 23 languages (12 more than Audio8-TTS-Preview), Apache-2.0 on Hugging Face.","api":"none","openWeights":"yes","params":{"text":"601M (601,159,424 in the main model, excluding the audio codec)","source":"Hugging Face model card and safetensors metadata","total":601159424},"facts":[{"label":"Base model","value":"Edge0/Audio8-TTS-Preview-0.6b (11 languages)"},{"label":"Languages","value":"23: Arabic, Cantonese, Chinese, Dutch, English, Finnish, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swedish, Thai, Turkish, Vietnamese"},{"label":"Architecture","value":"DualAR (inspired by Fish Audio S2 Pro): a 24-layer slow AR predicts one semantic token per frame; a 4-layer fast AR predicts 10 codec codebooks (4,096 entries each); width 896"},{"label":"Codec","value":"Bundled 44.1 kHz neural audio codec, about 21.5 frames per second; context up to 2,048 text/audio positions"},{"label":"Voice cloning","value":"Zero-shot from a reference recording plus its exact transcript"},{"label":"Weights","value":"Hugging Face (Edge0/Audio8-TTS-Realistic-0.6b), custom Transformers code (trust_remote_code)"},{"label":"License","value":"Apache-2.0"},{"label":"Serving","value":"SGLang Omni adapter with an OpenAI-compatible speech API; free Hugging Face Space playground"},{"label":"Paper","value":"arXiv 2610.14776, first posted Oct 10, 2026 (UTC); the weights were announced on Oct 29, 2026"}],"claims":["Lab claim: currently #1 on the Hugging Face TTS Arena (the model card says only that it is participating and has received encouraging ratings)."],"sources":[{"kind":"lab","title":"Samuel Zeng (Edge0 founder) announcement on X","url":"https://x.com/SamuelZengML/status/2115881491645554870","date":"2026-10-29"},{"kind":"lab","title":"Edge0 release links on X","url":"https://x.com/SamuelZengML/status/2115881504102551856","date":"2026-10-29"},{"kind":"repo","title":"Edge0/Audio8-TTS-Realistic-0.6b on Hugging Face","url":"https://huggingface.co/Edge0/Audio8-TTS-Realistic-0.6b"},{"kind":"repo","title":"Edge0-AI/Audio8_TTS on GitHub","url":"https://github.com/Edge0-AI/Audio8_TTS"},{"kind":"lab","title":"Audio8-TTS-Realistic-0.6B samples","url":"https://edge0-ai.github.io/Audio8_TTS/realistic-0.6b/"},{"kind":"paper","title":"Training a Compact, Strong, and Realistic Text-to-Speech Model (arXiv 2610.14776)","url":"https://arxiv.org/abs/2610.14776"}],"firstBenchmark":null},"organization":{"id":"edge0","name":"Edge0","slug":"edge0","country":null,"url":null,"color":"#A39E8F","mark":"ED"},"price":null,"answer":{"word":"Pending benchmarks.","scope":"Announced Oct 29, 2026; no independent results yet.","lead":"As of Nov 3, 2026, Audio8-TTS-Realistic-0.6B is pending benchmarks on isitsota: announced Oct 29, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}