Google launches Gemma 4 12B for local AI on laptops (opens in new tab)

Google introduced Gemma 4 12B, an open-weights AI model designed to run multimodal workloads locally on laptops with 16GB of VRAM or unified memory. Google DeepMind said the 11.95-billion-parameter model sits between its smaller E2B and E4B models and larger Gemma 4 models, supports text, image and native audio input, and is released under an Apache 2.0 license. The company described Gemma 4 12B as a unified, encoder-free multimodal model, with visual and audio inputs feeding into the languag...