Tencent has released Hy-SimulMT, an open-source AI model designed for simultaneous Chinese-English interpretation. The system continuously processes incoming speech and produces translated output, with support for both text and speech.

Across coverage, outlets emphasize the same core capabilities while highlighting implementation details. TechNode describes the release as including model weights and inference software, and notes real-time voice cloning intended to retain the original speaker’s vocal characteristics. It also points readers to Tencent’s Hy Translate app as a way to try the technology.

Pandaily similarly focuses on Hy-SimulMT as a speech-to-speech interpreter that uses the speaker’s own voice, and it adds a performance framing: Tencent’s stated long-form latency is about 3 seconds. Together, the reports frame the release as both an open model release and an interpretion product that targets low-delay, speaker-preserving translation.