Anthropic brings free Claude research credits to 20 scientific partners

Free AI access is the immediate offer; new scientific discoveries remain the ambition. Anthropic announced on October 28, 2026 that its AI for Science program now involves 20 scientific foundations, institutions and research groups worldwide. The expansion provides free Claude credits for research, though the company says many collaborations are still in their early stages.

Three routes from credits to researchers

The program began in May 2025, supporting individual scientists working on high-impact projects. Anthropic describes the expansion as a shift toward institutional partnerships and collaborative research. Scientists have told the company they use Claude to develop hypotheses, plan experiments, analyze complex data and communicate their findings.

The credits will reach researchers through three routes. Science funders and societies will distribute them to researchers; institutions will use them within their own organizations; and specific research initiatives will receive credits directly. That structure gives both organizations supporting outside scientists and teams conducting their own research a role in the expansion.

The funder and society group includes the Alfred P. Sloan Foundation, Simons Foundation, Schmidt Sciences, the US National Science Foundation and CAS, a division of the American Chemical Society. It also includes Renaissance Philanthropy’s Open Source for Science Fund, which supports the open-source software underlying scientific work.

Institutional recipients span astronomy and life sciences. The Space Telescope Science Institute runs science operations for NASA’s Hubble, James Webb and Roman telescopes. Other partners include the Wellcome Sanger Institute, Francis Crick Institute, Imperial College London’s Department of Life Sciences, European Molecular Biology Laboratory and Gladstone Institutes.

Brain maps, climate models and computer-checked proofs

Direct project support targets problems with potential public impact, rather than a single scientific discipline. The newly listed initiatives show how different that work can be:

• FlyWire at Princeton produced the first complete wiring map of an adult fruit fly’s brain. It develops open data and tools for studying how neural circuits produce perception and behavior.
• Columbia University’s LEAP combines machine learning with physics to improve how climate models represent clouds, oceans and land.
• AstroPath at Johns Hopkins applies calibration methods from astronomical surveys to tumor imaging, seeking consistent cell-by-cell measurements across slides and laboratories.
• Tau Ceti is building a shared mathematics library in Lean, a language that allows computers to check proofs.

Two other initiatives focus on genetic resources. The Broad Institute’s gnomAD pools DNA data from more than 800,000 people into an open reference used in rare-disease research and diagnosis. Australia’s Centre for Population Genomics works to include multicultural communities in genomic resources and builds tools for rare-disease diagnosis.

Existing collaborations show the technical direction

Anthropic also highlighted earlier collaborations to illustrate the program’s direction. With HHMI Janelia Research Campus, it developed the Model Hardware Standard, a shared specification intended to let AI agents operate laboratory instruments safely. The standard remains in research preview.

At the Allen Institute, which creates open datasets and tools for brain, cell and immune-system research, teams of AI agents are being built to analyze those data. Anthropic says that work turns months of manual analysis into hours. That speed claim concerns this collaboration, not the expanded program as a whole.