# DeepMind Reveals Gemini Robotics 2.5 and Open Safety Benchmark "ASIMOV-jailbreak"

> **Platform:** [SPIDITS AI](https://spidits.com/) — Real-Time AI News & Market Intelligence  
> **Published:** 2026-11-10T15:00:39.000Z  
> **Category:** PRODUCT_LAUNCH  
> **Impact Score:** 170/100  
> **Primary Source:** [Ars Technica](https://arstechnica.com/google/2026/11/deepmind-reveals-gemini-robotics-2-5-and-open-safety-benchmark-asimov-jailbreak)  
> **Canonical Citation:** [https://spidits.com/timeline/deepmind-reveals-gemini-robotics-2-5-and-open-safety-benchmark-asimov](https://spidits.com/timeline/deepmind-reveals-gemini-robotics-2-5-and-open-safety-benchmark-asimov)

## Executive Summary
The new model is up to twice as successful at some tasks.

## Why It Matters (Strategic Analysis)
The deployment demonstrates the transition of multimodal foundation models into physical embodied systems capable of handling deformable objects like ribbons and industrial kitting. It crystallizes the architectural trade-off between flexible generalized intelligence and the rigid 99.9 percent precision required by mission-critical factory automation.

## Key Entities & Companies
- **Google DeepMind**

## Referenced Coverage & Sources
- **[Ars Technica](https://arstechnica.com/google/2026/11/deepmind-reveals-gemini-robotics-2-5-and-open-safety-benchmark-asimov-jailbreak)**: DeepMind reveals Gemini Robotics 2.5 and open safety benchmark "ASIMOV-jailbreak" — _The new model is up to twice as successful at some tasks._

---
*Synthesized by SPIDITS AI Market Intelligence Desk. Track live AI news, model releases, and funding: [https://spidits.com](https://spidits.com)*