Do you understand what Google just did?
They put a frontier AI model on your laptop. free. open source.
it's called Gemma 4 12B. and it just dropped.
here's what makes it wild:
→ runs locally with just 16GB of VRAM
→ vision, audio, and agentic reasoning built in
→ performance nearing Google's much larger models
→ no separate encoders: one unified architecture handles everything
→ apache 2.0 license: use it for anything, including commercial projects
this isn't a stripped-down demo model.
it's a full multimodal AI that sees, hears, reasons, and acts.
on your laptop.
without the cloud.
without the cost.
without sending your data anywhere.
the architecture change is the most interesting part.
most multimodal models bolt on separate vision and audio encoders. Gemma 4 12B removes them entirely. one unified model handles every modality from the ground up.
the result: more capable. smaller footprint. faster locally.
显示更多
Today we’re introducing Gemma 4 12B — our latest open model that brings advanced agentic reasoning, vision and audio directly to your laptop.
It delivers performance nearing our larger Gemma models with a much smaller total memory footprint, while being small enough to run locally with just 16GB of VRAM. It’s open and accessible for everyone to use under a permissive Apache 2.0 license.
This is all made possible by our new, unified architecture that removes separate multimodal encoders. Here’s how we did it 🧵
显示更多