Yun Zhisheng announced that the company recently fully completed the U2-ASR and U2-TTS capability upgrades. U2-ASR added 13 new international language recognition capabilities, U2-TTS added 8 Southeast Asian language speech synthesis capabilities, and already supports over 100 Chinese dialects and more than 15 international languages. Under the unified evaluation scale, the U2-ASR has an average typographical error rate of 6.58% in 113 languages, and won 1st place in the limited data plan and 2nd place in the open data plan in the ChinaVoices Challenge 2026 Chinese Multi-dialect Speech Recognition Challenge. The upgraded model has been fully launched on the TokenHub platform and the standard API has been opened.

Yun Zhisheng announced that the company recently fully completed the U2-ASR and U2-TTS capability upgrades. U2-ASR added 13 new international language recognition capabilities, U2-TTS added 8 Southeast Asian language speech synthesis capabilities, and already supports over 100 Chinese dialects and more than 15 international languages. Under the unified evaluation scale, the U2-ASR has an average typographical error rate of 6.58% in 113 languages, and won 1st place in the limited data plan and 2nd place in the open data plan in the ChinaVoices Challenge 2026 Chinese Multi-dialect Speech Recognition Challenge. The upgraded model has been fully launched on the TokenHub platform and the standard API has been opened.