multilingual speech

Text-to-speech models that cover many languages: OmniVoice lists 600+, Higgs TTS 3 lists 100+, Fish Audio S2 Pro lists 80+, MagpieTTS Multilingual 357M v2607 lists 12 and Qwen3-TTS lists 10.

Related coverage

• NVIDIA releases MagpieTTS Multilingual 357M v2607, 12 languages with preset voices 2026-07-21
• Boson AI releases Higgs TTS 3, a 4B zero-shot voice cloning model for 100+ languages 2026-06-04
• k2-fsa releases OmniVoice, a 0.6B zero-shot TTS covering 600+ languages 2026-03-30
• Mistral releases Voxtral TTS, a 4B open-weights speech model, non-commercial 2026-03-26
• Fish Audio releases S2 Pro, a 5B TTS model for 80+ languages, research licence 2026-03-09
• Qwen releases Qwen3-TTS, Apache 2.0 voice cloning in 10 languages 2026-01-22
• FunAudioLLM releases Fun-CosyVoice3-0.5B-2512, an Apache 2.0 voice cloning TTS 2025-12-11