Evolution
Three generations, still evolving
AI Assistant Entry
System-level wake-up and voice interaction wired into settings and everyday apps — so "just say it" becomes the baseline experience.
Multimodal · Cross-app
Vision understanding combined with screen semantics enables cross-app task execution and continuous dialogue — evolving from understanding to seeing and getting it done.
Device-Cloud Collaborative Agent
On-device model scheduling working with cloud LLMs, with personalized memory and context awareness — it understands what should happen right now.
