Google Gemma

Google's Gemma is a family of lightweight, state-of-the-art open language models built on the same research and technology that powers Google's flagship Gemini models. Available in 2B and 7B parameter versions — with both base and instruction-tuned variants — Gemma is designed for developers, researchers, and enterprises who want powerful language model capabilities without the infrastructure overhead. Built with Google's AI safety principles at its core, Gemma offers cross-device compatibility and is optimized for Google Cloud and NVIDIA GPUs. It's freely accessible on Kaggle and Google Colab, and can also be deployed through Google Cloud with Vertex AI or GKE, making it one of the most accessible high-quality open models available today.