Anthropic has introduced Claude for Life Sciences, a package of product updates and partnerships aimed at making its AI assistant more useful across scientific research, clinical work, and regulatory tasks.
The company said the effort is meant to move Claude beyond isolated jobs such as summarizing papers or writing analysis code. Its longer-term goal is to support more of the life sciences workflow, from early discovery through commercialization. Anthropic said that increasing the pace of scientific progress is part of its public benefit mission.
At the center of the announcement is an improvement to Anthropic’s flagship model, Claude Sonnet 4.5. The company said the model performs better than earlier versions on life sciences benchmarks, including tests of laboratory protocol understanding and bioinformatics work. Anthropic compared Sonnet 4.5’s score on a protocol QA benchmark with both a human baseline and its predecessor, Sonnet 4, and said the newer model also improved on a separate bioinformatics evaluation.
Anthropic is pairing the model upgrade with tools intended to make Claude more practical in scientific settings. Those include new connectors to outside platforms, support for Agent Skills, a prompt library tailored to life sciences, and dedicated help from subject matter experts.
The new connectors are designed to let Claude interact directly with specialized research tools and data sources. Anthropic said the additions include Benchling, BioRender, PubMed, Scholar Gateway from Wiley, Synapse.org, and 10x Genomics.
According to the company, these integrations give Claude access to lab records, scientific figures, biomedical literature, peer-reviewed research, collaborative data projects, and single-cell and spatial analysis tools. Anthropic said the connectors are meant to make it easier for researchers to ask questions in natural language and receive results tied back to source material.
The life sciences connectors build on existing integrations with workplace software and data platforms, including Google Workspace, Microsoft SharePoint, OneDrive, Outlook, Teams, Databricks, and Snowflake.
Anthropic also pointed to its recently announced Agent Skills feature as a fit for scientific work. Skills bundle instructions, scripts, and resources that Claude can use to handle specific tasks in a more consistent way.
The company said it has started building life sciences-oriented skills, beginning with a tool for quality control and filtering of single-cell RNA sequencing data. Anthropic said researchers will also be able to create their own custom skills.
Anthropic says Claude can now support a range of tasks, including literature reviews, hypothesis generation, protocol drafting, bioinformatics analysis, and regulatory compliance work.
The company said Claude can summarize biomedical studies, help create study protocols and consent documents through the Benchling connection, process genomic data with Claude Code, and assemble outputs in slides, documents, or notebook formats. It also said the system can help draft and review regulatory submissions and compile compliance data.
To help users get started, Anthropic is also publishing a library of prompts for common life sciences tasks.
Anthropic is backing the rollout with hands-on support from its applied AI and customer teams, along with partnerships with consulting and technology firms such as Deloitte, Accenture, KPMG, PwC, Quantium, Slalom, Tribe AI, and Turing. AWS and Google Cloud are also part of the effort.
The company highlighted existing customers and partners, including Sanofi, Benchling, the Broad Institute of MIT and Harvard, 10x Genomics, Genmab, Komodo Health, Novo Nordisk, Stanford University, Schrödinger, and others. Their comments emphasized uses ranging from document automation and analytics to AI agents for research workflows.
Anthropic said its AI for Science program will continue to provide free API credits to selected researchers working on high-impact projects. The company is still accepting submissions for new ideas.
The launch positions Claude more directly against other AI systems being adapted for scientific research, while signaling Anthropic’s push to turn its model into a broader assistant for regulated, data-heavy industries.