Anthropic has opened a call for applications focused on rare genetic diseases within its AI for Science research support program. Accepted applicants receive up to 50,000 USD (about 8.1 million yen) in Claude credits over six months. The program runs on two tracks, one for basic science researchers and one for early-stage biotechs working on drug development, and applications close on August 2, 2026.

※1 USD = 162 JPY (as of July 21, 2026)

Rare Disease Chosen for the First Thematic Call

AI for Science launched in May 2025 as a program that supports scientific research through access to Anthropic's API[2]. Since then it has backed projects ranging from drug repurposing to quantum simulation[1].

Running the program, Anthropic found that projects tend to be more generative when several grantees work on related questions and trade tips with each other. That observation led the company to introduce thematic calls within the broader program, and rare genetic diseases were chosen as the first theme[1].

Rare diseases involve small numbers of patients individually, but in aggregate they are among the most prevalent conditions on the planet. An estimated 400 million people live with one of more than 7,000 rare diseases[1]. Because those patients are scattered across small populations, building patient registries, identifying promising therapeutic targets, and designing clinical trials all become difficult. Each condition is also typically defined by unique features such as a specific genetic variation or combination of symptoms, so they are studied in isolation and shared mechanisms across diseases are nearly impossible to spot. On top of that, rare diseases face the same challenge as all drug development: the time it takes to move promising candidates into patient trials.

Anthropic adds a footnote acknowledging that some sources place the number of rare diseases as high as 10,000, and that there is no agreed-upon definition of what a rare disease even is. Terminologies such as Orphanet, OMIM, GARD, and ICD each define disease differently, with some excluding chromosomal disorders, some ignoring conditions with environmental causes, and some requiring classification under a single anatomical system[1].

Track One: Basic Science and Data Infrastructure

The first track aims to foster collaboration between clinical researchers, patient organizations, and data scientists in order to speed up basic science and the discovery of mechanisms underlying rare diseases[1].

An early partner is the Monarch Initiative, an international consortium working to improve diagnosis and mechanism discovery for rare disease patients. Monarch develops standards and resources including the Mondo Disease Ontology, a computational framework and coding system that reconciles disease definitions scattered across OMIM, Orphanet, ICD, and dozens of other sources, and the Monarch Knowledge Graph, which integrates genotype-phenotype data across species.

More recently, Monarch contributors have been assembling data and knowledge into a new agent-friendly mechanistic disease classification library called DisMech. There, Claude can read case reports, variant databases, registry schemas, and raw public data, and point out mechanistic similarities between diseases at a pace and scale that has not been possible before. Monarch is inviting the new grantees to use and contribute to resources such as Mondo and DisMech[1].

Example projects for track one look like this. Propose and rank mechanistic links between distinct rare diseases that share a gene or pathway, presenting candidate relationships with evidence an expert can validate in DisMech. Curate and summarize patient organization data to conduct or improve natural history studies. Build evaluations that measure how well models handle rare disease tasks such as revealing candidate mechanisms for variants of unknown significance, phenotype-to-disease matching, and mechanism prediction, including an honest accounting of where they fail. Outputs from this track will be made publicly available at Monarchinitiative.org, and the program will be augmented by community-building efforts such as future rare disease hackathons[1].

Track Two: Compressing Drug Development Timelines

The second track supports biotechnologists and early-stage biotechs accelerating drug development for rare diseases. Today it takes one to two years to move from a confirmed genetic diagnosis to a treatment available to patients. Much of that time goes to waiting in queues for certified manufacturing slots, running safety studies sequentially instead of in parallel, and hand-assembling the thousands of pages of chemistry and regulatory documentation required for in-patient testing[1].

Anthropic believes parts of this process can be radically compressed with Claude. The main target is documentation work, such as drafting and reviewing the regulatory dossier. Beyond that, the company points to therapeutic strategy selection, where a model analyzes whether a target is druggable across modalities including small molecules, antibodies, and genetic medicines, and to the search for shared mechanisms across individual genetic therapies. If shared mechanisms are found, those therapies could be approved under a single basket trial instead of requiring a separate IND for each patient[1].

Anthropic also cites work already underway. Every Cure, an existing AI for Science grantee, uses Claude to identify drug repurposing opportunities across millions of candidates. The Centre for Population Genomics, a collaboration between the Garvan Institute and the Murdoch Children's Research Institute, is building a Claude-based system that drafts variant classifications for expert review, addressing one of the biggest bottlenecks in diagnosing rare genetic conditions. The Violet Research Institute, a small nonprofit researching ultra-rare genetic diseases affecting fewer than 1 in 50,000 births, uses Claude to navigate FDA guidelines, run bioinformatics pipelines, analyze experimental data, and draft regulatory filings[1].

How to Apply, and the Limits Anthropic Acknowledges

Applications are accepted through August 2, 2026 at 11:59 PM PST. Accepted applicants can spend their credits on Claude Opus and other generally available models approved for use in biology. Projects that may run up against Anthropic's bio classifiers may be eligible for exemptions[1].

Anthropic frames the program as part of its work on beneficial deployments, extending the benefits of AI to areas that might not emerge naturally through market forces. At the same time, it is candid about the limits. Claude may shorten therapeutic development timelines and curate biological data more efficiently than human teams alone, but it cannot help where the data is too paltry or too poorly organized for agents to reach. It may also struggle with parts of the diagnostic odyssey that involve insurance authorization or access to diagnostic facilities and infrastructure. Rare disease, Anthropic says, is simply too big a problem for one organization or one approach, and the company hopes efforts by other institutions to generate more high-quality longitudinal data, along with robust public-private partnerships, will complement this program[1].

Summary

Anthropic's AI for Science program has opened a call for applications centered on rare genetic diseases, offering accepted applicants up to 50,000 USD in Claude credits over six months. Track one works with the Monarch Initiative's Mondo and DisMech resources to advance basic science and mechanism discovery, while track two targets the time-consuming stages of drug development such as regulatory drafting and therapeutic strategy selection. Applications close on August 2, 2026. With Anthropic itself drawing a clear line around what AI cannot reach in data-poor areas and institutional barriers, the next six months should show how much of rare disease research these tools can actually move.

Source [1] https://www.anthropic.com/news/rare-disease-research-grants

Source [2] https://www.anthropic.com/news/ai-for-science-program