Edge0 releases Audio8-TTS-Realistic, a 0.6B open voice cloning TTS model

Edge0 has released Audio8-TTS-Realistic, an open-weight text-to-speech model that clones a voice from a single reference clip and transcript. The 601M-parameter model supports 23 languages, with weights and code under Apache 2.0. It reports 4.86% average CV3 error across 9 languages versus 7.73% for Qwen3-TTS-1.7B.

• 601M parameters, voice cloning from 1 clip and transcript
• Supports 23 languages, 12 more than its predecessor
• 4.86% average CV3 error across 9 languages vs 7.73% for Qwen3-TTS-1.7B
• Weights and code under Apache 2.0; 1.1–1.2 GiB peak on Apple M2