Anthropic LSVP Opens Claude Models to Verified Biology Research
Anthropic opens beta access to less restrictive Claude biology models, with verified grants, offline monitoring, 30-day retention and major launch limits.
Summary
Anthropic announced the beta Life Sciences Verification Program, LSVP, in a post dated September 17, 2026, following a September 16 publication timestamp. After onboarding dozens of organizations, it opened applications and expects hundreds in the first week, then most of the life sciences community within weeks. Verified academic labs, startups, pharmaceutical companies and institutions gain less restrictive access for drug discovery, research biology, clinical development, manufacturing, quality, regulatory work, supply chains, investing and diligence across Claude Science, Claude.ai, Claude Code and the API. Reviews cover research credentials, security and ethical oversight. Annual Standard Use team grants cover most workflows on Mythos 5.1, Opus 5, Sonnet 5 and future models.
Six-month, project-specific High-risk Use add-ons remove all life sciences request blocks for Opus 5 and Sonnet 5. Mythos high-risk access remains limited to a small, additionally vetted group while Anthropic works with the US government; cyber and other safeguards remain. LSVP replaces per-request blocking with offline monitoring for account compromise, insider diversion and dangerous agent activity, checking traffic against declared, nonsensitive use cases and alerting administrators. Flagged data is retained for 30 days, excluded from training and inaccessible to Anthropic’s life sciences researchers. The beta supports Anthropic’s API console plus Claude Enterprise and Team, but not individual Pro or Max plans, third-party platforms or BAA-enabled organizations; PHI requires separate non-BAA, non-HIPAA organizations. API and Claude Science support grant switching, while Claude.ai and Claude Code initially use a preset grant, except Claude Code with API authentication. Xaira, Edison and Manifold Bio plan to apply the access to medicine discovery.
Positives
- Dozens of organizations joined early access, with hundreds expected during the first week and broader life sciences coverage planned within weeks.
- Annual Standard Use grants unlock Mythos 5.1, Opus 5 and Sonnet 5 for most team-based biology workflows.
- Six-month High-risk Use grants remove all life sciences request blocks for approved Opus 5 and Sonnet 5 projects.
- Thirty-day flagged data is compartmentalized, barred from model training and inaccessible to Anthropic’s life sciences research teams.
- Xaira, Edison and Manifold Bio plan to use Anthropic’s models in drug and medicine discovery systems.
Risks & concerns
- Mythos high-risk access remains restricted to a small group undergoing additional vetting while Anthropic works with the US government.
- Thirty-day data retention is mandatory for flagged LSVP activity because offline monitoring replaces real-time request blocking.
- BAA-enabled organizations cannot join the beta, forcing PHI users into separate non-BAA, non-HIPAA environments.
- Account takeovers, malicious insiders and autonomous agents remain key threats when safeguards become more permissive.
- Individual Pro and Max plans and third-party platforms are unsupported at launch.
- Claude.ai and Claude Code initially limit users to one preset grant, except when Claude Code uses API authentication.