Understand this story
On November 10, 2026, Google DeepMind released the vision-language-action (VLA) model Gemini Robotics 2.5, calling it the most capable VLA model to date. According to information published by Google DeepMind, the model improved success rates by 30% to 100% over the previous baseline across 200 tasks, and increased task completion speed by 20% to 30%. In proof-of-concept trials at Mercedes-Benz factories in Hungary and Germany, the Apptronik Apollo 2 humanoid robot equipped with the model autonomously sorted automotive parts, achieving a 96% success rate. The model is being rolled out to robotics partners, and the open-source benchmark ASIMOV-Jailbreak was also released, on which Gemini 4 Argon achieved a 99.5% jailbreak resistance rate. A follow-up report by 36Kr·AI added that the success rate for the cake-picking task rose from 5% to 100%, and for untying bows from 0% to 70%, and stated that the model's development was accelerated by Gemini 4 Argon. The model is currently in the rollout phase to robotics partners.
AI-generated from public reports
Developments
1 confirmed developments
- Source:Google DeepMind
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 12
heise online
Nvidia CEO Jensen Huang has been revealed to have another office romance, and ByteDance also plans to launch an AI agent similar to Meta Muse
Nvidia CEO Jensen Huang has been revealed to have a relationship with Kimberly Powell, the vice president in charge of the healthcare business, in addition to his relationship with partner Alice Huang. The company's stock price continues to fall, and the board of directors will hold an emergency meeting to discuss countermeasures. ByteDance plans to release an AI agent similar to Meta Muse in the coming weeks that can handle daily tasks such as emails and online ordering. Co-founder Zhang Yiming said he hopes to rival Anthropic and OpenAI within three to four years. Brazil's Supreme Electoral Court also fined Meta about 1.2 million euros for failing to comply with court orders related to the election campaign.
Nov 11
36Kr · AISelected
Google DeepMind Releases Gemini Robotics 2.5 Robot Model
Google DeepMind has released the VLA robot model Gemini Robotics 2.5, which improves the success rate by 30% to 100% over the baseline across 200 tasks and increases completion speed by 20% to 30%. Among these, the success rate for the cake-picking task rose from 5% to 100%, and for bow-untying from 0% to 70%. In proof-of-concept trials at Mercedes-Benz plants in Sindelfingen, Germany, and Kecskemét, Hungary, the Apollo 2 humanoid robot equipped with this model achieved a 96% success rate in sorting automotive parts. The model was developed with accelerated support from Gemini 4 Argon, which can reconstruct simulation environments from RGB photos and analyze the causes of failures, and has a jailbreak resistance rate of 99.5% on the ASIMOV-Jailbreak benchmark.
heise online
Google DeepMind releases Gemini Robotics 2.5; Mercedes-Benz pilot achieves 96% success rate in sorting metal parts
Google DeepMind has released the vision-language-action model Gemini Robotics 2.5, saying that across more than 200 tasks its success rate improved by 30% to 100% compared with 2.0, and it completes tasks 20% to 30% faster. In a pilot with Apptronik and Mercedes-Benz, the Apollo-2 humanoid robot completed the sorting and sequencing of automotive metal brackets with a 96% success rate, but robotics lead Kanishka Rao acknowledged that there is still a long way to go to reach near-human reliability. DeepMind also released ASIMOV-Jailbreak, an open-source safety benchmark of about 1,500 prompts, now available on Hugging Face.
Huxiu
Google DeepMind Releases Gemini Robotics 2.5, Achieving a 96% Sorting Success Rate at Mercedes-Benz Factory
On November 10, Google DeepMind released Gemini Robotics 2.5, calling it its strongest robot vision-language-action (VLA) model to date, with a relative success rate improvement of 30% to 100% across more than 200 tasks and a 20% to 30% increase in execution speed. In proof-of-concept trials on Mercedes-Benz production lines in Germany and Hungary, the Apollo 2 humanoid robot equipped with the model completed parts sorting, grasping, and placement with a task success rate of 96%. The article also mentions that Gemini 4 Argon has been used to generate simulated training scenarios and analyze failure causes, and compares Claude Fable 5.2's 64.9% success rate on 42 tasks in the RoboDojo Sim benchmark with GPT-6 Astra's 29.8%.
The PaperSelected
Google DeepMind Releases VLA Model Gemini Robotics 2.5
Google DeepMind released on local time November 10 the new-generation vision-language-action model Gemini Robotics 2.5, which greatly improved the success rate of dexterous manipulation in evaluations of about 200 tasks; for example, using tongs to pick up pastries rose from 5% to 100%, and untying a ribbon bow rose from 0% to 70%. The model completes tasks 20%-30% faster than the previous version, and performance on multi-step and contact-intensive tasks also improved. The team also conducted a proof of concept with Apptronik on Mercedes-Benz production lines in Hungary and Germany, where the humanoid robot Apollo 2 achieved a 96% success rate in picking up and placing automotive metal brackets.
Zhidongxi
Google DeepMind Releases Gemini Robotics 2.5, Achieving a 96% Success Rate in Real-World Testing on Mercedes Production Lines
Google DeepMind has released its next-generation VLA model, Gemini Robotics 2.5, claiming that across tests on more than 200 tasks, its success rate improved by a relative 30% to 100%, and its execution speed increased by 20% to 30%. In an automotive parts sorting PoC in collaboration with Apptronik and Mercedes, the Apollo 2 equipped with this model achieved a 96% task success rate. The model also introduces an inference-time guidance mechanism, improving the success rates of safety-related tasks near obstacles and near people by 33% and 43%, respectively.
Nov 10
Testing Catalog
Google Releases Gemini Robotics 2.5 Physical AI Model
Google has released Gemini Robotics 2.5, a new model for physical AI, built with the support of the Gemini 4 Argon model. In a proof of concept with Apptronik and Mercedes-Benz, Gemini Robotics 2.5 helped the Apollo 2 robot pick and organize automotive parts for assembly with 96% accuracy.
Google DeepMindFirst-partySelected
Google DeepMind Releases Gemini Robotics 2.5 Vision-Language-Action Model
Google DeepMind has released Gemini Robotics 2.5, calling it the most capable vision-language-action (VLA) model to date, with success rates improved by 30% to 100% over the previous baseline across 200 tasks and task completion speed improved by 20% to 30%. In proof-of-concept trials at Mercedes-Benz factories in Hungary and Germany, the Apptronik Apollo 2 humanoid robot equipped with the model autonomously sorted automotive parts, achieving a 96% success rate. The model is being rolled out to robotics partners, alongside the release of the open-source benchmark ASIMOV-Jailbreak, on which Gemini 4 Argon achieved a 99.5% jailbreak resistance rate.
Heat trend for this story
- Current
- 5
- Period peak
- 97 Nov 11 21:00
- 24-hour change
- -90%
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.