Google DeepMind announced Gemini Robotics 2.5 on November 10, 2026 — its most capable vision-language-action (VLA) model to date. The launch included results from a live pilot on Mercedes-Benz production lines in Kecskemét, Hungary, and Sindelfingen, Germany, where Apptronik's Apollo 2 humanoid robot achieved a 96% success rate sorting metal automotive brackets, according to data published by Google DeepMind.
What Google DeepMind Announced with Gemini Robotics 2.5
The system runs on a two-model architecture: Gemini Robotics ER handles reasoning and overall task flow, while Gemini Robotics 2.5 controls precise arm and hand movements. According to Google DeepMind, the system was evaluated across more than 200 distinct tasks and achieves 30%–100% improvement over previous baselines depending on task type. The model is rolling out exclusively to select partners — Apptronik, Agile Robots, and Boston Dynamics — with no public API or published pricing. DeepMind's own robotics team acknowledged that scaling from the current 96% to the 99.9% reliability standard common in industrial manufacturing will be the real challenge ahead.
"The move of AI models from the desktop to the factory floor is one of the clearest signals of where automation is heading. SMBs that understand these advances today can position themselves ahead of the curve when these systems reach the broader market."
Davarion Group & LabsWhat Changes for SMBs
- 01No direct access yet: Gemini Robotics 2.5 is reserved for Google's robotics partners; there is no public API or pricing at this time.
- 02Manufacturing SMBs should start preparing now: document repetitive processes and map bottlenecks in assembly or quality inspection.
- 03Software automation already available today — AI agents, automated workflows — is the practical first step before considering physical robotics.
- 04By 2027, evaluate robotics-as-a-service providers that may integrate these models as they move toward broader market availability, based on current adoption timelines.
The Gemini Robotics 2.5 announcement confirms that the gap between software AI and physical automation is closing quickly. For SMBs with manufacturing or logistics operations, this is the moment to audit processes that are currently manual and repetitive, and to begin building the data infrastructure that future robotic systems will require.
At Davarion Group & Labs, we help Houston SMBs assess their AI automation readiness, starting with the tools available today: software agents, automated workflows, and process analysis. If your business has repetitive operations that currently depend on manual work, we can help you map the path forward.
