San Francisco, September 18 — Anthropic has opened applications for a beta program that gives life science professionals access to its Mythos, Opus and Sonnet models under a refined set of safeguards built to be more permissive for biology-related work. The company announced the Life Sciences Verification Program, or LSVP, on September 17. It follows an early-access phase in which Anthropic said it had already onboarded dozens of organizations, and it is now being extended to the broader life science community.
The program launches in beta, initially for teams and institutions, with plans to expand access to individual Pro and Max plans over time.
What the program unlocks
The LSVP is designed to let life science professionals use the models for tasks currently blocked in Anthropic’s generally available Fable models, including drug discovery, research biology, clinical development and manufacturing. It is aimed at teams of all kinds, from academic labs to startups and pharma companies. To qualify, each applicant goes through a verification process reviewing research credentials, security standards and ethical research oversight.
Verified teams may then apply for two types of LSVP grants: Standard Use or High-risk Use. Grants can be used across all product surfaces, including Claude Science, Claude.ai, Claude Code and the API.
Standard Use grants are described as suitable for most life science work, including the majority of biology research and development workflows, and can be extended to entire teams for diverse daily workloads. They are renewed once a year and give teams access to the Mythos, Opus and Sonnet models with refined classifiers more permissive for science tasks than generally available models. Standard Use grants apply to Mythos 5.1, Opus 5 and Sonnet 5 today, and to future models as they launch.
They are designed to enable activities spanning basic science, R&D, supply chain and manufacturing, clinical development, quality assurance, regulatory affairs, investing and diligence, and more.
A second tier for higher-risk work
High-risk Use is an add-on grant for teams working in areas blocked under Standard Use. It removes all safeguards that block life sciences requests, applies to a single research project rather than a full team, and must be renewed every six months. Anthropic gave the example of a single researcher with dual-use work holding one Standard Use grant for daily activities and one or more High-risk Use grants for specific projects, such as characterizing how one specific family of viral vectors is recognized by human immune pathways.
High-risk grants for Claude Opus 5 and Claude Sonnet 5 are available today. The company said it is working with the US government to make high-risk grants more broadly available for Claude Mythos, but at launch they remain limited to a small set of entities with additional vetting.
All other safeguards, such as cyber classifiers, remain in place under LSVP grants.
Why the safeguards look different
Anthropic cited its recent threat report, saying there are increasingly sophisticated misuse attempts on its platform, including attempts that could support biological weapons development. In biology, it said, it is often not possible to differentiate between a user doing valid work — for example, researching a viral pathogen to develop vaccines against it — and pursuing harm, for example trying to increase the transmissibility of a virus maliciously.
The most concerning threat models are ones where valid access has been diverted or overtaken by an actor with bad intent; insider threats and rogue use have been major factors in significant biosafety incidents and scares. Anthropic said it aimed to protect against three concerning threat models in particular.
In close collaboration with enterprise CISOs, Anthropic designed the LSVP safeguards around shared responsibility, monitoring usage against the intended use case for the model access. Because it vets LSVP organizations for life sciences credibility and oversight, it empowers them to specify what constitutes safe usage for teams or projects within their program. Each entity’s access is tied to the use cases it specified in its grant applications, and Anthropic continuously monitors LSVP traffic to identify usage or patterns outside the stated safe scope.
Should unauthorized activity occur, it can flag cases to organization admins to take action within pre-agreed timeframes for triaging and remediating incidents. Use cases should include high-level descriptions of intended work, like one would share in a job listing, and not include sensitive information or IP.
Anthropic said serious misuse is often spread across many requests and sessions to look disconnected and evade detection. In the LSVP, it is shifting safeguards from real-time blocking, where potentially harmful access is rejected at the time of each request, to offline monitoring, which allows clearer identification of potential misuse across patterns of behavior. Shifting enforcement from real-time blocking to offline monitoring allows legitimate work to proceed with fewer interruptions but requires retaining data associated with flagged activity for review.
For LSVP traffic, Anthropic is requiring data retention for 30 days to do this monitoring effectively. This data is strictly compartmentalized and cannot be used for model training or accessed by members of Anthropic’s life sciences research teams.
For organizations that qualify, Anthropic said it is also working on additional
