Monday 16 November
Global

Your region, kept on this device.

DeepMind unveils Gemini Robotics 2.5 and an open robot safety benchmark

Severity
Info
Sources
1

has released , the vision-language-action component of its physical AI platform, with as the planning vision-language layer. The model was evaluated across more than 200 activities, including pipetting and deformable-object manipulation, and was demonstrated on 's humanoid. The article cites no CVEs or affected versions; the security-relevant item is the open ASIMOV-jailbreak safety benchmark for testing embodied AI behaviour.

For security analysts · Written by AI from 1 source · Wrong?

What to do

  1. No action needed.

Written by AI from the sources, checked against them

Full story timeline · 1 update

  1. Tue 10 NovemberDeepMind reveals Gemini Robotics 2.5 and open safety benchmark "ASIMOV-jailbreak"Ars Technica (all)

Who reported · 1 outlet

Ars Technica (all)

Who and what is in it

Report a problem