Qwen3-Omni and Qwen3-TTS Released Simultaneously!

The Qwen team has released two new models: Qwen3-TTS and Qwen3-Omni. Qwen3-TTS-Flash supports 17 voice styles and 10 languages, including multiple international languages and Chinese dialects, with latency as low as 97 milliseconds, delivering outstanding performance. Qwen3-Omni comes in three versions—an all-modal description model, a standard edition, and a reasoning edition—supporting interaction in 119 text languages, speech understanding in 19 languages, and speech generation in 10 languages.