商汤发布对 GPT-6-Astra 具身智能的实证研究，覆盖 12 个仿真基准和 8 个真实机器人平台，并开源 SenseNova-RoboRSI 框架及 27B 模型 RoboRSI-27B-Preview，其任务成本约为 GPT-6-Astra 的 1/50。在统一协议（ICL 加最多 10 倍交互预算）下，GPT-6-Astra 在 12 个基准上均达到 SOTA，RoboDojo 从 1 倍预算的 27.6% 升至 10 倍预算的 81.9%，DexJoCo 为 87.3%，高于此前 SOTA 的 40.3%。研究同时指出，即便把笔放进笔筒这样的简单任务，推理成本也约 55 美元，且真实机器人会过热、等待时间过长。
商汤在 12 个仿真基准和 8 个真实机器人平台上评测 GPT-6-Astra，并给出开源蒸馏模型与成本对比。
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.
We release the most comprehensive 𝗲𝗺𝗽𝗶𝗿𝗶𝗰𝗮𝗹 𝘀𝘁𝘂𝗱𝘆 𝗼𝗳 𝗚𝗣𝗧-𝟲-𝗔𝘀𝘁𝗿𝗮’𝘀 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 across 12 simulation benchmarks and 8 real-robot platforms, together with 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮-𝗥𝗼𝗯𝗼𝗥𝗦𝗜: an open-source harness framework and a 27B model that runs at ~1/50 the cost of GPT-6-Astra. Here's what we found: 🔹 𝗧𝗵𝗲 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗰𝗲𝗶𝗹𝗶𝗻𝗴 𝗶𝘀 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗺𝘂𝗰𝗵 𝗵𝗶𝗴𝗵𝗲𝗿. Under a unified protocol (ICL + up to 10× the interaction budget), GPT-6-Astra achieves clear SOTA across all 12 benchmarks: RoboDojo: 27.6% at 1× → 81.9% at 10×, DexJoCo: 87.3% vs. 40.3% prior SOTA, while the LIBERO benchmarks are pushed to saturation. 🔹 𝗧𝗼𝗼 𝗺𝘂𝗰𝗵 𝗯𝘂𝗰𝗸 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗯𝗮𝗻𝗴. Even a simple task like placing pens in a holder costs ~$55 in inference. 🔹 𝗧𝗵𝗲 𝘄𝗼𝗿𝗹𝗱 𝗱𝗼𝗲𝘀𝗻’𝘁 𝘄𝗮𝗶𝘁. Time matters as real robots overheat, and an egg burned while the model was still thinking. Our takeaway: GPT-6-Astra has largely crossed the threshold of 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗰𝗼𝗺𝗽𝗲𝘁𝗲𝗻𝗰𝗲, but 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗮𝗹 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 is not here yet. 🌐 Project page: https://www.sensenova.ai/has-gpt-6-astra-achieved-embodied-intelligence 📄 Report: https://www.sensenova.ai/gpt-6/assets/docs/has_gpt-6_achieved_embodied_intelligence.pdf 💻 Code & model: https://github.com/OpenSenseNova/SenseNova-RoboRSI在 X 查看被引用的帖子
来源：Rohan Paul · x.com