Introducing Gemma 4 12B: a unified, encoder-free multimodal model
Source
deepmind.google
Date
Context
Google DeepMind’s Gemma 4 12B puts more of image and audio processing inside the LLMA large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.Full definition → itself. In the June 2026 announcement, the company explains that it replaced separate multimodalA model that works with more than text — reading images, audio, or video, and sometimes generating them too.Full definition → encoders with lightweight input transformations: images pass through a small embeddingA list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.Full definition → module, while raw audio is projected into the representation used by the model. The intended benefit is a smaller memory footprint for local multimodal applications. Google describes the model as able to run with 16GB of video or unified memory, connecting this architecture to its goal of running image and audio applications on laptops.
Terms in this piece · Glossary
LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
multimodal — A model that works with more than text — reading images, audio, or video, and sometimes generating them too.
embedding — A list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.
