Chinese Researchers Release Unified Arabic Speech Model

Chinese researchers at Shanghai Jiao Tong University's X-LANCE Lab released Habibi, an open-source text-to-speech model published on arXiv in January 2026. Habibi unifies more than 20 Arabic dialects and supports zero-shot voice cloning from short reference audio clips, enabling rapid on-the-fly synthesis across dialects. The paper is unreviewed, and analysts say the release could strengthen China's technological presence in the Middle East.