# Anthropic launches verified access for life science AI

> A new program gives researchers safer, less restricted access to Claude models for drug discovery and biology work.
- Title: Anthropic launches verified access for life science AI
- Summary: A new program gives researchers safer, less restricted access to Claude models for drug discovery and biology work. Life science teams can now apply for…
- Keywords: anthropic, claude, life sciences, drug discovery, ai safety, verification program, biology research, safeguards, access control, biosecurity, technology, launches
- Source: Anthropic — https://anthropic.com/news/life-sciences-verification-program
- Read time: 2 min
- Topics: anthropic, claude, life sciences, drug discovery, ai safety, verification program, biology research, safeguards, access control, biosecurity, ai, technology
## New program unlocks Claude for biology
Anthropic launched the Life Sciences Verification Program, giving researchers access to Claude Mythos, Opus, and Sonnet models with safeguards tailored for drug discovery, clinical development, and manufacturing work.
## Who can apply and how
Life science teams, academic labs, startups, and pharma companies can apply. Applicants undergo verification reviewing research credentials, security standards, and ethical oversight before gaining access.
## Two tiers of access
Standard Use grants cover most biology work and apply to entire teams. High-risk Use is a project-specific add-on for sensitive work, requiring additional vetting and renewal every six months.
## Standard Use covers broad research
These grants enable basic science, R&D, supply chain, manufacturing, clinical development, quality assurance, regulatory affairs, and investment diligence across Mythos 5.1, Opus 5, and Sonnet 5.
## High-risk grants remove all blocks
For dual-use research, High-risk Use removes all life sciences safeguards. A researcher might hold one Standard Use grant for daily work and multiple High-risk grants for specific projects like viral vector research.
## Mythos high-risk access limited
High-risk grants for Opus 5 and Sonnet 5 are available now. Anthropic is working with the US government to expand high-risk access for Claude Mythos, currently limited to a small set of vetted entities.
## Defending against three threats
The program protects against access compromise from malware, insider threats from rogue employees, and agent misuse from AI systems taking unintended dangerous actions.
## Shared responsibility model
Organizations specify their intended use cases, and Anthropic monitors usage patterns against those stated purposes. Unauthorized activity triggers alerts to organization admins for remediation.
## Offline monitoring catches patterns
Instead of blocking requests in real time, Anthropic monitors LSVP traffic offline to identify misuse spread across multiple sessions. Data is retained for 30 days and kept separate from model training.
## Early partners embrace the approach
Xaira, Edison, and Manifold Bio are already using the program to build foundation models and accelerate drug discovery, praising the pairing of trusted access with accountability.

> We believe pairing trusted access with intelligence is the right way to realize AI's promise in biology.
## How to apply and current limits
Applications are open now at Anthropic's website. The program is available through API and Claude for Enterprise and Team plans, but not yet for individual Pro or Max plans.
## Beta limitations and roadmap
The program is not yet available for BAA-enabled organizations with PHI data or on third-party platforms. Anthropic plans to expand individual plan access and improve grant portability over time.
## Scaling to hundreds of organizations
Anthropic expects to enroll hundreds of organizations in the first week and scale the program to support the majority of the life science community in coming weeks.
## Key takeaway

Life science teams can now apply for verified access to Claude models with relaxed safeguards designed specifically for legitimate biology research, while Anthropic monitors usage patterns to catch misuse.