Models & Labs
Google DeepMind Unveils Gemini Robotics 2
WIRED AIhigh confidence
Why it matters
- →Gemini Robotics 2 integrates multiple AI models for advanced robotic control.
- →It highlights Google's focus on extending AI capabilities into the physical world.
- →Safety concerns are addressed with the new ASIMOV-Agentic benchmark.
Google DeepMind has launched Gemini Robotics 2, a new AI model designed to control various robots, including humanoids capable of complex tasks. The system integrates several AI models, allowing robots to understand and interact with their surroundings. Demonstrations showed robots autonomously performing tasks like tidying shelves. While this advancement highlights Google's expertise in robotics, it also raises safety concerns, leading to the introduction of the ASIMOV-Agentic benchmark for safety assessment. This release signals Google's push to expand AI's role from digital to physical applications.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
More from WIRED AI
More in Models & Labs
Models & Labsother
llama.cpp b12340 adds CUDA 13 and ROCm 10.0
This release quietly closes the hardware gap for local inference by adding default support for CUDA 13 and ROCm 10.0 across Linux and Windows. Users with newer NVIDIA GPUs or AMD accelerators no longer need to compile from source to get performance, while Apple Silicon builds now ship KleidiAI disabled by default, shifting the focus back to standard ARM optimizations. The inclusion of Snapdragon support for Linux and Android marks a significant step toward on-device AI beyond the x86/ARM desktop ecosystem. It is less about new features and more about ensuring llama.cpp remains the universal runtime for every available accelerator.
llama.cpp Releases
Models & Labscoding
llama.cpp b12342 adds few-row MMQ for speculative decoding
Speculative decoding just got a significant speed boost on NVIDIA GPUs. This release introduces a specialized kernel for 'few-row' matrix multiplications, targeting the small batch sizes typical of verification passes in speculative decoding workflows. By optimizing these specific operations for quantized formats like Q8_0 and IQ4 variants, llama.cpp reduces latency where it matters most during token generation. The change is CUDA-only for now but leverages hardware capabilities from Ampere onwards, making local inference faster without requiring new model architectures or complex configuration changes.
