Anthropic's Life Sciences Program Brings High-Risk Grants & 30-Day Monitoring

Anthropic has opened its Life Sciences Verification Program (LSVP) to the broader life sciences community, giving verified organizations access to Claude models for biology work that may be blocked on its generally available models.

The beta program also introduces a high-risk access tier with additional vetting and monitoring.
 

TL;DR

 

• Anthropic’s LSVP gives verified life sciences teams broader access to Mythos, Opus, and Sonnet models for drug discovery, research, clinical development, manufacturing, and related work.
• Standard Use grants are renewed annually, while High-risk Use grants apply to individual projects and require renewal every six months.
• LSVP replaces some real-time blocking with offline monitoring and requires 30-day data retention for flagged activity.

 

Anthropic Life Sciences Verification Program Expands Claude Access For Drug Discovery And R&D

Anthropic opened applications for its Life Sciences Verification Program (LSVP) after already onboarding dozens of organizations through early access. The program is initially aimed at teams and institutions, including academic labs, startups, pharmaceutical companies, and other life sciences organizations.

Applicants undergo verification covering research credentials, security standards, and ethical research oversight before they can receive access. Verified organizations can then apply for Standard Use or High-risk Use grants depending on the work they need Claude to perform.

Standard Use is designed to cover most life sciences workflows and can be extended across an entire team. It currently supports Mythos 5.1, Opus 5, and Sonnet 5, with Anthropic saying future models will also be included as they launch.

That access spans areas including basic science, R&D, clinical development, manufacturing, supply chains, quality assurance, regulatory affairs, investing, and diligence.
 

 

High-Risk Claude Grants Remove Life Sciences Blocks With Additional Vetting

For projects that go beyond Standard Use limits, Anthropic is introducing High-risk Use grants that remove safeguards specifically blocking life sciences requests. Unlike Standard Use, these grants apply to individual research projects rather than whole teams and must be renewed every six months.

High-risk access is currently available for Claude Opus 5 and Sonnet 5. Anthropic said it is working with the US government to broaden similar access for Mythos, although Mythos high-risk grants will initially remain limited to a small number of additionally vetted organizations.

Other safeguards remain active, including Anthropic’s cyber classifiers.

The company says the structure is built around “shared responsibility,” with verified organizations defining the intended scope of their work while Anthropic monitors activity for behavior that falls outside those approved use cases.
 

Anthropic Uses 30-Day Offline Monitoring To Catch Life Sciences Misuse

The program also changes how Anthropic handles potentially risky biology-related activity.

Instead of relying entirely on real-time blocking for every potentially harmful request, LSVP uses offline monitoring designed to spot misuse spread across multiple requests or sessions. Anthropic identified access compromise, insider threats, and misuse by AI agents as three threat models the system is designed to address.

Because of this monitoring model, data associated with LSVP traffic must be retained for 30 days. Anthropic says the data is compartmentalized, cannot be used for model training, and cannot be accessed by members of its life sciences research teams.

Organizations only need to provide high-level descriptions of intended use cases and are instructed not to include sensitive information or intellectual property.
 

 

Claude Life Sciences Access Launches For Enterprise And Team Plans First

LSVP is available through Anthropic’s first-party API console and Claude Enterprise and Team plans, with applications now open. Anthropic expects to enroll hundreds of organizations during the first week before expanding the program further across the life sciences community.

Individual plans and third-party platforms are not supported at launch. Anthropic also said the beta is unavailable to BAA-enabled organizations, meaning customers working with PHI must use separate non-BAA organizations with non-HIPAA data.

The approach has already attracted life sciences companies including Xaira Therapeutics, Edison Scientific, and Manifold Bio. Xaira said it believes “pairing trusted access with intelligence is the right way to realize AI’s promise in biology.”

Anthropic says it plans to expand LSVP access to individual Pro and Max plans over time, alongside additional products, research collaborations, and improvements to the program.