Xiaomi AI Glasses × SenseNova: Creating a Portable Gateway to AI

In June 2025, Xiaomi launched its first AI glasses, positioned as "the next-generation personal smart device" and "a portable gateway to AI”.

This product serves as a real-time AI assistant that perceives the world, which is its key differentiation from traditional smart glasses. Its market reception has been strong: according to China’s XR Vision reports, nearly 50,000 units were likely sold within the first three days of release, setting a new record for the fastest-selling AI glasses in China.

As Xiaomi's long-term ecosystem partner, SenseNova large model interaction platform SenseChat has been experimentally integrated with Xiaomi AI glasses. Functioning as a "second brain” for users, this integration enables seamless support across real-life scenarios in work and daily life, empowering users through “sight, speech, memory and processing” all in one device.

Full-duplex real-time interaction, stable and smooth human-computer interaction

Traditional smart devices often fall short in delivering natural and interactive experiences due to response delays, contextual disconnects, and one-way input mechanisms.

One of the standout features of the Xiaomi AI Glasses × SenseTime SenseNova collaboration is the full-duplex real-time audio-visual interaction capability. This goes beyond simple voice command chains or half-duplex Q&A exchanges, instead creating a natural, fluid, and uninterrupted interactive loop akin to human conversation. For instance, when visiting stores or exhibitions, the glasses provide real-time recognition and act as an interactive encyclopedia, while serving as a translation assistant on travels. These scenarios demonstrate the system's robust capabilities:

Millisecond-level Response, Communication at the Speed of Thought: Leveraging SenseNova's robust model inference capabilities and underlying optimizations, interaction latency is reduced to milliseconds. The system begins comprehension before the user finishes speaking, generating nearreal-time responses.

Unbroken Contextual Continuity: The SenseNova large model deeply understands context, accurately tracks dialogue flow, and supports natural interruptions, corrections, and follow-up questions, making conversations as smooth as interacting with a human assistant.

Exceptional Stability in Complex Environments: Even in noisy exhibition halls or bustling streets, the upgraded noise suppression ensures commands are captured precisely and understood accurately.

Deep Analysis with Memory Augmentation: By integrating audio-visual memory and retrieval technology, the system instantly recalls and references past interactions (e.g., when reviewing proposals with clients) to deliver highly targeted information support.

Jointly Exploring New Paradigms for Next-Generation Smart Hardware Interaction

Through their collaboration, SenseTime and Xiaomi AI Glasses are exploring how large models can drive both technological advancement and user experience innovation in smart devices. SenseTime’s SenseNova’s interactive capabilities have been significantly enhanced following its recent upgrade to V6.5. Its multimodal interaction performance now surpasses Gemini 2.5 Flash and GPT-4o, demonstrating outstanding advancements across all metrics.

This collaboration provides valuable insights for smart glasses and the broader next-generation smart hardware interaction paradigms, enabling AI to integrate more deeply and in real-time into human perception and interaction, while continuously unlocking its potential to empower user-driven value creation.

Looking ahead, SenseTime is committed to push the boundaries of large model technologies for real-time audio-visual interaction and multimodal understanding, extending "efficient, natural, and deeply collaborative AI capabilities" across broader applications. These advancements will enable comprehensive support ranging from personal productivity enhancement to industry-level solutions.