Force Origin DM0.5 tops six embodied-AI benchmarks, sweeps all four RoboColiseum boards
机器之心 · wechat · 2026-09-16
RoboColiseum, an embodied-model evaluation launched by robot maker AgiBot, tests instruction following, spatial understanding, disturbance adaptation and general manipulation — and Force Origin's DM0.5 ranked first on all four boards, the only model to sweep the suite. It now leads all six public benchmarks it has entered, including 99.0% on LIBERO and 43.0% on the real-robot RoboChallengeTable30V2.
Key points:
- RoboDojo Memory dimension: 47.74 points vs 13.37 for second place
- Real-world demos: recovers from camera viewpoint shifts and human interruptions, millimeter-scale micro-block assembly with failure self-detection, force-aware cucumber slicing via motor current sensing, conveyor logistics at 3s per parcel with >99% accuracy
- Architecture natively supports 60s of history; pretraining spans six robot embodiments, 50k hours of real-robot manipulation data; inference latency cut 9.29x from 534ms to 57.49ms
- Fully open-sourced, already deployed in a major retailer's warehouse and a 3C manufacturer's factory
More from Embodied
- Wearing Whoop and Apple Watch on a single dual-slot band — tinyfool · 2026-11-02
- REA: async brain-cerebellum runtime for robots where policies never wait on the LLM — Pengfei Zhou · 2026-11-02
- Robot Combat Sports Coming to Japan with a Permanent Arena in Tokyo — cixliv · 2026-11-02
- REK robot fighting in the cage: humanoid bot in boxing gloves draws a crowd — broodsugar · 2026-11-02
- iPhone Duo: a folding iPhone mod impresses, but its hinge raises crackgate concerns — rudrank · 2026-11-02
- InflatableBots: University of Calgary's inflatable robots bring dynamic haptics to large-scale VR — lukas_m_ziegler · 2026-11-02