OpenAI's Dots agent is set to receive a key upgrade with the addition of the Terra model, which is both fast and low-cost. This model can break a single task into parts and automatically schedule the appropriate models to complete it in relay according to each step, replacing the previous approach of running a single model all the way through. While ensuring core reasoning performance, it greatly reduces task time and cost. This upgrade will push Dots out of its high-price exclusive positioning and toward more ordinary users. However, multi-model relay scheduling for long tasks is more difficult; OpenAI previously experienced system failures due to automatic model routing, and the actual implementation results of Dots still need to be verified through real-world testing. Currently, major technology companies are all laying out strategies in the personal agent track, and the core of competition has shifted from single flagship model benchmark scores to efficiently scheduling resources and stably delivering task results that require no rework.