Meta releases Muse Glimmer, a 30B open-source Apache 2.0 agentic model for local deployment
“an agent like that needs deep access to personal context”
on-device
8 tracked signals on on-device.
Google's Gemma 4 brings frontier-level, multimodal, agentic AI to offline edge and mobile devices.
“The North Star of our smaller models is intelligence per byte of memory footprint.”
Google releases EmbeddingGemma 2, a sub-billion open model unifying text, image, video, and audio embeddings on-device.
“a picture of a cat, the word cat, and the sound of a cat are all mapped close to each other in a shared high-dimensional embedding space”
NVIDIA Cosmos 3 Edge brings 4B world model robot control to on-device hardware
NVIDIA RTX Spark laptops run large AI agent models locally with 128GB unified memory, partnering with Microsoft for private on-device workflows.
“Because of its 128 GB of unified memory, it can run big models.”
Apple's MLX stack lets developers run full agentic AI loops locally on Mac with no cloud or API keys.
“No cloud, no API keys, just your hardware doing the work.”
Microsoft extends Foundry Local to run agentic AI fully disconnected for sovereign and enterprise environments.
“AI needs to be operated fully disconnected while still um delivering real time situational awareness and decision support.”
Gemma 4 runs 10 parallel local sub-agents at 170+ tokens/sec on one machine.
“This is the new standard for local parallel AI.”