从“会回答”到“会办事”，vivo如何解AI手机这道题？
Ranking
No observed public metrics; popularity remains neutral/archived.
Merged summary
TL;DR - vivo unveiled an agent-oriented AI phone architecture combining a device-cloud model matrix with a system-level Harness to turn user intent into cross-app, cross-device actions. The approach matters because it shifts differentiation from model benchmarks toward personalized context, orchestration, privacy-aware memory, and reliable execution.
- BlueLM-Nano handles on-device perception and memory, while BlueLM-Realtime supports speech; cloud models BlueLM-Flash and BlueLM-Pro cover fast task execution and complex reasoning.
- The Harness provides shared perception, memory, planning, and execution layers, connecting agents and services through 6,000+ system tools, MCP, A2A, CLI, and unified APIs.
- With user authorization, the system learns from behavior and task outcomes, using idle-device reflection to retain execution experience and improve future workflows.
- Partnerships with Alipay, Meituan, Amap, and JD.com demonstrate agent-driven flows spanning local services, travel, navigation, shopping, and payments.
Sources (1)
从“会回答”到“会办事”，vivo如何解AI手机这道题？
TL;DR - vivo unveiled an agent-oriented AI phone architecture combining a device-cloud model matrix with a system-level Harness to turn user intent into cross-app, cross-device actions. The approach matters because it shifts differentiation from model benchmarks toward personalized context, orchestration, privacy-aware memory, and reliable execution.
- BlueLM-Nano handles on-device perception and memory, while BlueLM-Realtime supports speech; cloud models BlueLM-Flash and BlueLM-Pro cover fast task execution and complex reasoning.
- The Harness provides shared perception, memory, planning, and execution layers, connecting agents and services through 6,000+ system tools, MCP, A2A, CLI, and unified APIs.
- With user authorization, the system learns from behavior and task outcomes, using idle-device reflection to retain execution experience and improve future workflows.
- Partnerships with Alipay, Meituan, Amap, and JD.com demonstrate agent-driven flows spanning local services, travel, navigation, shopping, and payments.