商汤开源 RoboRSI-27B-Preview：将具身经验内化，任务成本降至约 1/50
Why we picked it
商汤把 GPT-6-Astra 从演示里学到的具身经验，蒸馏进开源的 RoboRSI-27B-Preview，还把任务成本压到约 1/50，做机器人的可以看看。
SenseNova-RoboRSI 的核心思路是经验内化：GPT-6-Astra 通过 ICL 从演示中提取视觉示例、机器人状态、动作轨迹、工具和恢复策略等信息。商汤将这类具身经验蒸馏到开源模型 RoboRSI-27B-Preview 中。该模型具备主动感知、空间推理、规划与基于反馈的修正能力，任务成本约为原来的 1/50。
Original · rohanpaul_ai
The key idea behind SenseNova-RoboRSI is experience internalization.
GPT-6-Astra uses ICL to extract useful information from demonstrations, including visual examples, robot states, action trajectories, tools, and recovery strategies.
SenseTime distills this embodied experience into the open-source RoboRSI-27B-Preview, which demonstrates active perception, spatial reasoning, planning, and feedback-driven revision at roughly 1/50 the task cost.
