Google launches Gemini for Science tools for researchers

Google has launched Gemini for Science, a package of experimental AI tools for researchers and science teams.

The rollout begins with three prototypes in Google Labs and a separate bundle, Science Skills, for life science research workflows.

The Labs tools are Hypothesis Generation, Computational Discovery and Literature Insights. They are intended to help researchers generate and test ideas, review published studies and organise complex information across large bodies of scientific material.

Hypothesis Generation is built with Co-Scientist. It works with researchers to define a research challenge, then uses multiple AI agents to generate, debate and evaluate hypotheses, with citations attached to support claims.

Computational Discovery is built with AlphaEvolve and Empirical Research Assistance, or ERA. Google said it generates and scores thousands of code variations in parallel, allowing scientists to test modelling approaches in areas such as solar forecasting and epidemiology.

Literature Insights uses Google NotebookLM. It searches scientific literature and structures the results into tables with searchable attributes for side-by-side analysis. Users can also query a curated body of papers through chat and create outputs such as reports, slide decks, infographics and audio or video overviews.

Science skills

Alongside those tools, Google introduced Science Skills, a bundle that integrates more than 30 life science databases and tools, including UniProt, AlphaFold Database, AlphaGenome API and InterPro.

Researchers can use Science Skills on agentic platforms including Google Antigravity to carry out workflows such as structural bioinformatics and genomic analysis more quickly, Google said. In one internal test, its team used the tools to complete an analysis in minutes rather than hours and identified potential mechanisms linked to a rare genetic disease involving mutations in the AK2 gene.

Research users

Some of the underlying tools are already in use in private preview by commercial and public sector organisations, according to Google. BASF is using AlphaEvolve to optimise supply chains, while Klarna is using it in machine learning work.

Daiichi Sankyo, Bayer Crop Science and the US National Labs, as part of the US Department of Energy's Genesis Mission, are also using Co-Scientist in research projects, the company said.

Several research papers tied to the tools have already been published, with the ERA and Co-Scientist papers appearing in Nature, according to Google.

Scientific partnerships

Google said it is working with more than 100 institutions to validate the systems. These collaborations include Stanford University on liver fibrosis, Imperial College London on antimicrobial resistance and a multi-year effort with the Crick Institute.

It has also built a tester community ranging from PhD students to industry researchers and Nobel laureates to assess the systems on complex real-world scientific problems.

In parallel, Google is developing tools for peer review and scientific validation with conferences including ICML, STOC and NeurIPS. It named the experimental Paper Assistant Tool and ScholarPeer among those efforts.

Broader push

The launch extends Google's push to position its AI systems as tools for scientific research rather than only consumer or office software. The company pointed to earlier work including AlphaFold, which it said has been used by more than 3 million researchers, and AlphaGenome, as well as products such as Google Scholar, Earth Engine, Colab, MedGemma, Earth AI and Gemini Deep Research.

Google framed the new package around the problem of information overload in modern science, where the volume of published work can make it difficult for individual researchers to connect findings across disciplines. It said the tools are intended to reduce time spent on manual synthesis, coding experiments and literature review.

Access to the new experiments will open gradually. Google said the offering is aimed at both individual researchers through Labs and organisations through Google Cloud.