Google has expanded its open AI portfolio with the launch of Gemma 4 12B. This runs directly on personal computers and workstations. The rollout allows users to access advanced AI tools without relying heavily on cloud support.
Built by Google DeepMind, this new model is powered by Google's Gemini tech. The tech giant claims that Gemma 4 12B hits that sweet spot between performance and hardware efficiency. Thus, helping developers create slick AI apps that regular hardware can handle.
Unlike large cloud-based models that require tons of computing resources, Gemma 4 12B performs well on consumer-grade GPUs and modern desktop systems. Given this, organizations and developers can cut down cloud costs while boosting privacy and speed.
Plus, it's released under the Apache 2.0 license, which means both commercial and research projects can use it. Google thinks it'll fit into many industry applications.
Gemma 4 12B also does better at reasoning and following instructions. It handles multi-step tasks and complex workflows way better than its predecessors. Besides, Google added features like function calling, structured JSON outputs, and support for system instructions. This lets developers create AI agents that can work with software tools, apps, and services straight from local devices.
Gemma 4 12B handles multiple types of input, like text, images, and video. Users can do things like image analysis and OCR right on this platform; there's no need to send data elsewhere. It helps with coding, debugging, and boosting developer productivity. Plus, it works with over 140 languages and can even handle long documents all in one go.
This release shows Google's dedication to delivering advanced AI locally to users, which helps protect privacy, reduces costs, and speeds up tasks.
Also read: Google Pixel Watch 5 Leak: Alleged Prototype Found Underwater Before Launch