Google Launches Gemma: Lightweight Open AI Models for Developers
Google introduces Gemma, a new family of state-of-the-art open models built from the same technology as Gemini. These lightweight models aim to help developers and researchers build AI responsibly.
Google introduces Gemma, a new family of lightweight, state-of-the-art open models built from the same research and technology used to create the Gemini models. Developed by Google DeepMind and other teams across the company, Gemma expands Google's long history of contributing open-source AI innovations like TensorFlow and BERT to the community. The model weights are available worldwide starting today.
The company releases model weights in two sizes, Gemma 2B and Gemma 7B, with both pre-trained and instruction-tuned variants available. Google also provides a new Responsible Generative AI Toolkit that offers guidance and essential tools for creating safer AI applications. Developers receive toolchains for inference and supervised fine-tuning across all major frameworks, including JAX, PyTorch, and TensorFlow through native Keras 3.0.
Pre-trained and instruction-tuned Gemma models run efficiently on laptops, workstations, or Google Cloud, with easy deployment options on Vertex AI and Google Kubernetes Engine. Optimization across multiple AI hardware platforms ensures industry-leading performance on NVIDIA GPUs and Google Cloud TPUs. Furthermore, the terms of use permit responsible commercial usage and distribution for all organizations, regardless of size.