On October 20, 2025, Anthropic announced "Claude for Life Sciences," a set of model improvements, integrations and tooling intended to support life-science work across the full research lifecycle — from early discovery through translation and commercialization.
Model improvements
Anthropic highlighted performance gains in its flagship model, Claude Sonnet 4.5. On the Protocol QA benchmark (multiple choice, 10-shot prompting) Sonnet 4.5 scored 0.83, compared with a human baseline of 0.79 and Sonnet 4’s 0.74. The company also reports similar improvements on BixBench, a bioinformatics evaluation.
New connectors and integrations
Claude is gaining several new connectors designed to let the model access experimental records, publications, figures and analytical platforms directly:
- Benchling: enables Claude to answer questions with links back to source experiments, notebooks and records;
- BioRender: connects to a vetted library of scientific figures, icons and templates;
- PubMed: access to millions of biomedical research articles and clinical studies;
- Scholar Gateway (developed by Wiley): provides authoritative, peer-reviewed scientific content inside Claude;
- Synapse.org: supports collaborative sharing and analysis of data in public or private projects;
- 10x Genomics: allows single-cell and spatial analyses via natural language.
These add to existing integrations with Google Workspace, Microsoft SharePoint, OneDrive, Outlook, Teams and platforms such as Databricks and Snowflake for large-scale bioinformatics analytics and dataset search.
Agent Skills and scientific workflows
Anthropic released Agent Skills — folders with instructions, scripts and resources that Claude can use to follow specific tasks predictably. Skills are positioned as a natural fit for scientific protocols and procedures. A notable initial skill is single-cell-rna-qc, which performs quality control and filtering on single-cell RNA-seq data following scverse best practices.
Scientists and organizations can also build custom skills for Claude; Anthropic provides guidance for setting these up.
Use cases in life sciences
Anthropic outlines typical applications of Claude for Life Sciences:
- Research: literature reviews, summarization and hypothesis generation, with Claude citing and summarizing biomedical literature and proposing testable ideas;
- Protocol generation: drafting study protocols, standard operating procedures and consent documents via the Benchling connector;
- Bioinformatics and data analysis: processing and analyzing genomic data with Claude Code and presenting results as slides, documents or code notebooks;
- Clinical and regulatory compliance: drafting and reviewing regulatory submissions and compiling compliance data.
To assist adoption, Anthropic is also creating a prompt library with templates for these tasks.
Partners, customers and support
Anthropic offers hands-on support through Applied AI and customer-facing teams and partners with consultancies that help organizations adopt AI in life sciences — including Caylent, Deloitte, Accenture, KPMG, PwC, Quantium, Slalom, Tribe AI and Turing — plus cloud partners AWS and Google Cloud.
The announcement cites a range of customers and collaborators already using Claude for scientific tasks, including Sanofi, Benchling, Broad Institute of MIT and Harvard, 10x Genomics, Genmab, Komodo Health, Novo Nordisk, Stanford University, PwC, Schrödinger, Latch Bio, EvolutionaryScale, Manifold, FutureHouse and Axiom Bio. Anthropic’s release includes partner statements highlighting efficiency gains, improved workflows and the potential to accelerate drug development and bioinformatics analysis.
AI for Science program and availability
Anthropic continues its AI for Science program, which provides free API credits to support leading researchers working on high-impact scientific projects. The company welcomes submissions for project ideas to help identify new applications for Claude.
Claude for Life Sciences is available via Claude.com and on the AWS Marketplace; Google Cloud Marketplace availability is noted as forthcoming. The Sonnet 4.5 Protocol QA result is documented in the Sonnet 4.5 System Card (see referenced pages for details). Anthropic states the combined model, connector and tooling updates aim to make Claude a more capable research partner across the life-science R&D pipeline.
