Rumor
Zhipu reportedly planning to release an open-source Text-to-Speech model.
Saturday, December 6, 2025 at 02:00 PM
Zhipu, an AI company, is rumored to be preparing for the launch of a new open-source Text-to-Speech (TTS) model in the near future.
Context
Z.ai (formerly Zhipu AI) is reportedly planning to release an open-source Text-to-Speech (TTS) model, a move consistent with its recent open-source multimodal roadmap. The company has already launched GLM-4-Voice, an end-to-end speech model capable of understanding and generating Chinese and English speech with emotional control, based on GLM-4-9B. This follows their rebranding to Z.ai and the release of GLM-4.5 and GLM-4.5 Air in mid-2025, demonstrating a rapid model development cadence.
Such a release is significant for investors as it would likely accelerate the adoption of Z.ai's technology in China and across Asia. An open-source TTS model could drive increased demand for inference compute, including GPUs and NPUs, and expand addressable markets for voice-enabled applications. The company also secured a 61.28 million yuan government deal in May 2025 and announced compatibility with Huawei Ascend processors in August 2025, further solidifying its market position and potential for growth in the AI supply chain.
Related Companies
Z
Z.ai