Google DeepMind is an AI research and development organisation focused on advancing artificial intelligence and building it responsibly to benefit humanity. Its model cards provide structured overviews of how its AI models are designed and evaluated, helping people learn about the systems and their development.

The organisation’s portfolio includes Gemini models for tasks such as intelligent agents, image creation and editing, and audio interaction. Its generative models include Veo for video, Lyria for music and audio, and Imagen for image generation. Gemini Robotics models are designed to help robots perceive, reason, use tools, and act in the real world. DeepMind also offers Gemma, a family of lightweight open models for building AI applications, including specialised options for tasks such as coding, embeddings, and image understanding.

Beyond models, Google DeepMind conducts research and develops AI tools for science, including AlphaFold for protein structure prediction, AlphaGenome for studying the human genome, and WeatherNext for forecasting. Its website also shares research publications, evaluations, safety information, educational initiatives, and news about its work.