DeepMind unveils Gemini Robotics 2.5 and an open safety benchmark called ASIMOV-jailbreak

Google DeepMind has announced Gemini Robotics 2.5, a new robotics model that reportedly performs up to twice as well on some tasks. The company also released an open safety benchmark named ASIMOV-jailbreak alongside the model. The benchmark is intended to test whether robotic systems can be manipulated into unsafe behavior. The source article provides few further details on the specific tasks, evaluation methods, or availability timeline.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in