Claude for Life Sciences

Source: Anthropic Date: 2025-10-20 URL: https://www.anthropic.com/news/claude-for-life-sciences

Summary

Anthropic launched Claude for Life Sciences on October 20, 2025, featuring Sonnet 4.5 scoring 0.83 on Protocol QA benchmarks (vs. 0.79 human baseline). New integrations include Benchling, BioRender, PubMed, Scholar Gateway, Synapse.org, and 10x Genomics. Agent Skills include single-cell-rna-qc for RNA sequencing analysis. Partners include Deloitte, Accenture, KPMG, PwC (consulting) plus AWS and Google Cloud (cloud), and Sanofi, Genmab, Novo Nordisk, and Stanford University (life sciences).

Implications

• Enterprise vertical thread. Life sciences joins financial services as a Anthropic industry vertical with dedicated capability packages, benchmarks, and partner ecosystems. The pattern: pick regulated, high-value industries, show domain benchmark leadership, sign consulting firms as implementation partners.
• Protocol QA benchmark (0.83 vs. 0.79 human). Outperforming human baseline on scientific protocol tasks is the kind of metric that gets cited in pharma procurement conversations. Worth tracking how this benchmark holds as competitors publish equivalents.
• Partner density. All four major consulting firms (Deloitte, Accenture, KPMG, PwC) plus hyperscalers (AWS, Google Cloud) plus named pharma/biotech customers is a fully mature vertical go-to-market. This didn’t exist six months earlier.
• Agent Skills specificity. single-cell-rna-qc as a named Agent Skill is more specific than anything previously shipped — suggests Anthropic is building deep enough to compete with specialized biotech AI vendors (Benchling AI, Recursion), not just general-purpose LLM providers.
• Watch: Sanofi/Genmab/Novo Nordisk deployment scale; FDA/EMA regulatory positioning; how the life sciences Agent Skills compare to vendor-specific tools in Benchling.