Gemma

Gemma is a family of open models from Google DeepMind for developers building AI applications. The models are designed to run across cloud servers, personal computers, laptops, mobile devices, and IoT devices.

The Gemma family includes models focused on advanced reasoning, mobile and memory efficiency, multimodal applications, translation, embeddings, medical imaging, function calling, and encoder-decoder workflows. The site provides model information, documentation, integrations, and access through Google AI Studio.

Highlights

• Open models for cloud, desktop, mobile, and IoT deployments
• Variants including Gemma 4, Gemma 3 270M, MedGemma, TranslateGemma, and FunctionGemma
• Models for reasoning, multimodal applications, translation, embeddings, and medical imaging
• Documentation, integrations, and model variants available on the product site