Anthropic's Life Sciences Verification Program: two grants, 30-day retention

Anthropic introduced the Life Sciences Verification Program (LSVP) on 17 September 2026, a beta that gives verified life-science organisations access to its Mythos, Opus and Sonnet models with safeguards that are, in Anthropic's words, more permissive for biology-related work. Anthropic says generally available Fable models block tasks such as drug discovery, research biology, clinical development and manufacturing; LSVP is its route around those blocks for vetted teams. The company says it had already onboarded dozens of organisations in early access and expected hundreds in the first week.

Two kinds of grant

• Standard Use: for most life-science work, including the majority of biology R&D workflows. It can cover entire teams for daily workloads and renews once a year. Anthropic says it applies to Mythos 5.1, Opus 5 and Sonnet 5 today, and to future models as they launch.
• High-risk Use: an add-on for work blocked under Standard Use. Anthropic says it removes all safeguards that block life-sciences requests, applies to a single research project rather than a team, and must be renewed every six months. Grants for Opus 5 and Sonnet 5 are available now; for Mythos they are limited to a small set of vetted entities while Anthropic works with the US government on wider availability.

Anthropic says other safeguards, such as cyber classifiers, stay in place under both grants. Applicants go through verification covering research credentials, security standards and ethical research oversight.

The trade: monitoring instead of blocking

Anthropic says LSVP moves enforcement from real-time blocking to offline monitoring. Each organisation's access is tied to the use cases it states in its application, described at the level of a job listing and without sensitive information or IP, and Anthropic says it continuously monitors traffic for activity outside that scope and flags it to organisation admins to act within pre-agreed timeframes.

To do that, Anthropic requires 30-day retention of LSVP data. It says this data is strictly compartmentalised and cannot be used for model training or accessed by its life-sciences research teams. It also says it is working on how qualifying organisations can combine LSVP with Enterprise Frontier Safeguards, which its 1 September post describes as zero-data-retention privacy with customer-controlled storage.

Where it does not work yet

• Surfaces: LSVP is available in Anthropic's first-party console for API use and in Claude for Enterprise and Team plans. Individual plans and third-party platforms are not supported yet.
• HIPAA: as a beta, LSVP is not available for organisations with a business associate agreement (BAA). Anthropic says customers with protected health information (PHI) should use separate non-BAA organisations.
• Switching grants: in the API and Claude Science users can switch between grants. In Claude.ai and Claude Code only a preselected default grant applies at first, except in Claude Code with API authentication.

What the post leaves out

The post is vendor-published, and the named endorsements come from Xaira Therapeutics, Edison Scientific and Manifold Bio, who say they are pleased to take part. It gives no pricing, no verification turnaround, no list of what the Standard Use classifiers still block, and no accuracy figures for the monitoring. Anthropic's 30 September changelog note clarifies that its linked form is an expression of interest, not an application, so budget time for a vetting step.

For a team in pharma or biotech, the practical checks are whether your data policy allows 30-day retention, whether your work needs a team-wide grant or a per-project one, and whether any PHI workflow sits in a BAA organisation that cannot join yet.