Credit: CFP

AsianFin — Xiaomi Corp. unveiled a new open-source AI voice model on Monday, stepping up efforts to integrate artificial intelligence into its automotive and smart home offerings as China ’ s tech giants vie to lead in multimodal AI applications.

The MiDashengLM-7B model, an evolution of Xiaomi ’ s in-house voice recognition system, has been optimized for deployment across vehicles and household devices. Developed with Alibaba Group Holding Ltd. ’ s open-source Qwen2.5-Omni-7B as a foundation, Xiaomi showcased its technical improvements and performance benchmarks in a WeChat post.

The Beijing-based smartphone and EV maker is racing to diversify beyond its core handset business, with electric vehicles now a top strategic priority. As part of this pivot, Xiaomi is banking on AI-driven ecosystems to enhance its product integration and user experience.

China ’ s tech sector has rapidly embraced open-sourcing AI models to build developer communities and accelerate adoption. Industry giants from Alibaba to Tencent Holdings Ltd. have launched their own models in recent months that span images, video, and audio, aiming to rival tools like OpenAI ’ s Sora.

AI has also become a geopolitical focus. Both U.S. President Donald Trump and Chinese President Xi Jinping have underscored the need for national AI leadership, intensifying the global race for technological supremacy.