Introducing the Life Sciences Verification Program

Anthropic has launched the Life Sciences Verification Program, a beta access track for life science professionals who want to use Mythos, Opus, and Sonnet on biology work that the company’s generally available models block. The pitch is straightforward: let the right people do more useful work, but make them prove who they are first. Applications are now open beyond the early-access group that already included dozens of organizations.

The first version is for teams and institutions, not individuals. Anthropic says it will expand to individual Pro and Max plans over time. For now, the program is meant to cover work across drug discovery, research biology, clinical development, and manufacturing, and it is open to organizations ranging from academic labs to startups and pharma companies.

Getting in starts with verification. Applicants go through a review of research credentials, security standards, and ethical research oversight, then qualify for one of two grants: Standard Use or High-risk Use. Standard Use is the broad option and can be renewed yearly. It applies to Mythos 5.1, Opus 5, and Sonnet 5 today, across Claude Science, Claude.ai, Claude Code, and the API, with access intended for the everyday mix of biology R&D, manufacturing, regulatory work, diligence, and the rest.

High-risk Use is narrower and more sensitive. It is an add-on for specific projects, must be renewed every six months, and removes the safeguards that block life sciences requests. Anthropic says those grants are already available for Opus 5 and Sonnet 5, while Mythos high-risk access is still limited to a small set of entities as the company works with the US government to broaden it.

The company is also changing how enforcement works. Instead of blocking risky requests in real time, LSVP leans on offline monitoring, with 30 days of retained data for review. Anthropic says the aim is to catch patterns that spread across sessions and requests, including access compromise, insider threats, and agent misuse, while still letting legitimate work move with fewer interruptions. The tradeoff is obvious: more access, more oversight, and a lot more trust in the organizations that sign up.