Xiaomi Suddenly Releases It’s Own Free AI Model That Beats ChatGPT and Gemini

Xiaomi has introduced a new artificial intelligence (AI) MiMo-V2-Pro which has around one trillion parameters and easily challenge leading US AI systems while operating at a significantly lower cost.

According to available details, the model achieves performance levels that approach OpenAI and Anthropic and needs substantially few resources per interaction. It is also designed to transmit under 256,000 tokens per session.

The project is led by Fuli Luo, who was previously involved in the DeepSeek R1 initiative.

MiMo-V2-Pro & Omni & TTS is out. Our first full-stack model family built truly for the Agent era. I call this a quiet ambush — not because we planned it, but because the shift from Chat to Agent paradigm happened so fast, even we barely believed it. Somewhere in between was a process that was thrilling, painful, and fascinating all at once.

Rather than focusing solely on traditional text-based interaction, Xiaomi is steering the model toward executing actions and handling tasks in more autonomous environments.

“We will open-source — when the models are stable enough to deserve it,” Luo added.

The company, already a major player in consumer electronics and IoT systems, has expanded into new industries in recent years. It is among the world’s largest smartphone makers and has also entered the electric vehicle market with models such as the SU7 and YU7 SUV.