Gemini Robotics 2.5 hits 96% kitting parts on Mercedes lines
DeepMind's new robot model lifts task success 30–100%, and Apptronik's Apollo 2 used it to sort car parts on live Mercedes lines.
Google DeepMind's Gemini Robotics 2.5 has already been on live car lines. In a pilot with Apptronik and Mercedes-Benz, an Apollo 2 humanoid running the model did automotive kitting at Kecskemét, Hungary, and Sindelfingen, Germany. It picked metal brackets out of bulk bins and placed them into sequencing trays in the order each vehicle needs, with a 96% success rate. DeepMind calls it a proof of concept, not a production rollout.
The model is a vision-language-action (VLA) model. In the pilot, Gemini Robotics ER tracked the workflow while 2.5 controlled the hands and arms. Across 200 tasks it scores a 30–100% higher success rate than Gemini Robotics 2 and finishes 20–30% faster. Packing pastries with tongs went from 5% to 100%. DeepMind says Gemini 4 Argon sped up development by acting as a simulation engine that turns one real camera scene into thousands of practice runs. Apptronik CEO Jeff Cardenas says a few weeks of post-training carried over to the factory with very little drop-off.
On safety, reasoning at run time raised safe task completion by 33% around obstacles and 43% around nearby people. DeepMind is also open-sourcing ASIMOV-Jailbreak, a benchmark of disguised unsafe requests for robots. For now the model is rolling out to robotics partners only.
Why it matters: humanoids are moving from lab demos to measured results on real factory floors, and simulation is shrinking the time it takes to train them.
Sources
Independent coverage
- Gemini Robotics 2.5 accelerates real-world general dexterity deepmind.google
- Ziegler reports Gemini Robotics 2.5 use at Mercedes-Benz humanoid.guide
Written by an AI pipeline from the sources above. Methodology · Report an error
The stories that matter for people who build with AI, with the source for every claim. Free, no spam.
Feed, daily deep-dive and bytes — readable offline, with push alerts for the topics you follow.