DeepMind unveils Gemini Robotics 2.5: robots tie bows autonomously, 96% pick-and-pack success at Mercedes-Benz
AI_Andrew · x · 2026-11-13
Google DeepMind introduced Gemini Robotics 2.5, its most capable vision-language-action model, demoed live at CoRL 2026. Key points:
- Dexterity: tested on 200+ tasks; robots autonomously use tools like tongs and tie a bow on a box.
- Reliability: in a real proof-of-concept on Mercedes-Benz production lines, Apptronik's Apollo 2 humanoid picked and packed auto parts with a 96% success rate.
- Safety: improved physical-safety steerability plus a new benchmark for robotics-specific adversarial jailbreaks.
- Built on progress accelerated by Gemini 4 Argon, aiming to move robots into unstructured real-world environments.
More from Embodied
- SteerAct lifts robot action-model success ~10% with just 24 rollouts per task — pliang279 · 2026-11-13
- Deep tech cohort demos for investors: autonomous chemistry labs, underwater robots and 10 more startups — Scobleizer · 2026-11-13
- Rodney Brooks' Robust AI deploys Carter, a collaborative warehouse robot built around humans — MIT News AI · 2026-11-13
- CoreWeave Q3 revenue hits $3.66B, up 168%; signs multi-year Wayve training deal — thursdai_pod · 2026-11-13
- NVIDIA's BirdDreams jointly generates perspective and BEV video for driving world models — nvidia · 2026-11-13
- Galaxea Co-founder Hang Zhao's CoRL Keynote: Split Robot Intelligence into System 2/1/0 by Timescale — YuXiang_IRVL · 2026-11-13