Event ·Multi-source

商汤开源 RoboRSI-27B-Preview：将具身经验内化，任务成本降至约 1/50

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。

Current read

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。

9 sources47° AI heatLast updated 2026/10/23 15:32:59

Evidence

9 sources

Primary sourcerohanpaul_ai

2026/10/23 15:32:59View source

Related source 1Artificial Analysis

2026/10/21 15:53:56View source

Related source 2Teortaxes

2026/10/22 14:00:18View source

Related source 3Bindu Reddy

2026/10/23 11:06:01View source

Related source 4SiliconFlowAI

2026/10/23 13:42:34View source

Related source 5PolymarketMoney

2026/10/21 15:49:50View source

Related source 6arXiv cs.AI

2026/10/21 17:54:01View source

Related source 7TestingCatalog

2026/10/21 18:57:16View source

Related source 8Nous Research

2026/10/21 20:50:20View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story