Google DeepMind launched Gemini for Science, a suite of experimental research tools now live on Google Labs. The collection includes three prototypes: Hypothesis Generation, built with Co-Scientist for multi-agent idea tournaments; Computational Discovery, powered by AlphaEvolve and ERA to test thousands of code variations in parallel; and Literature Insights, leveraging NotebookLM for structured literature review. Each tool targets a core bottleneck in modern research.

Science Skills, bundled with Google Antigravity, integrates insights from over 30 life science databases including UniProt, AlphaFold Database, and AlphaGenome API. Early testing showed a complex genomic analysis that normally takes hours completed in minutes, revealing novel insights into a rare genetic disease tied to AK2 mutations. These tools arrive as the scientific community grapples with an ever-growing body of published research.

Enterprise partners like BASF, Klarna, Daiichi Sankyo, Bayer Crop Science, and U.S. National Labs are already running these tools in private preview. Research papers for ERA and Co-Scientist were published in Nature today. Google is also collaborating with over 100 institutions, including Stanford and Imperial College London, to validate the systems against real-world research challenges.