Google DeepMind unveils Gemini Robotics 2.5 after Mercedes pilot
Original title: Google DeepMind's Gemini Robotics 2.5 scores 96 percent sorting metal parts in a Mercedes-Benz assembly pilot
DeepMind reports 96 percent success sorting parts and is rolling the model out to Agile, Apptronik and Mercedes-Benz.
Google DeepMind introduced Gemini Robotics 2.5, a vision-language-action model that turns visual input and language instructions into robot movement. In pilots with Apptronik and Mercedes-Benz in Kecskemét and Sindelfingen, an Apollo 2 humanoid sorted metal parts into assembly trays. DeepMind reports a 96 percent success rate for that test and higher, faster results than its own earlier baselines across more than 200 tasks. It also published the open-source ASIMOV-Jailbreak benchmark.
Unconfirmed: The 96 percent figure and benchmark scores are Google DeepMind's own reported results.
- DeepMind says success rose 30 to 100 percent and speed 20 to 30 percent versus its baselines.
- Safe task completion improved 33 percent near obstacles and 43 percent near people.
- Gemini 4 Argon scored 99.5 percent jailbreak resistance, versus 82 percent for Gemini Robotics ER 2.
- The model was also tested on the Franka Duo platform.
- Vendors
- Google DeepMindApptronik
- Products
- Apollo 2Franka Duo
- Organizations
- Mercedes-BenzAgile
From the original article
2 min read at the-decoder.com
Google Deepmind introduced Gemini Robotics 2. 5, a new control model for robots, and tested it with Apptronik on active Mercedes-Benz production lines in Hungary and Germany. The humanoid robot Apollo 2 sorted metal parts there with a reported success rate of 96 percent.
By Maximilian Schreiner · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.