Anthropic loosens biology safeguards for verified life-science labs
On 17 September Anthropic opened a beta Life Sciences Verification Program that gives verified teams Mythos, Opus and Sonnet with more permissive biology safeguards than its generally available Fable models. Applicants face credential, security and ethics review, with Standard Use and High-risk Use grants. Newsquawk the same day reported the beta for teams and institutions and named Standard-use grants for Mythos 5.1, Opus 5 and Sonnet 5.
Artificial Intelligence··Night
A verified lane around biology blocks
Anthropic says the LSVP gives life-science professionals access to Mythos, Opus and Sonnet with safeguards more permissive for biology-related work than its generally available Fable models. It has already onboarded dozens of organizations in early access and is now opening applications, launching in beta first for teams and institutions, with individual Pro and Max plans later. Anthropic opened a beta Life Sciences Verification Program: verified teams get Mythos, Opus and Sonnet with classifiers it says are more permissive for biology work than its generally available Fable models. The company says it has already onboarded dozens of organizations in early access and is opening applications more widely, with individual Pro and Max plans later. A Standard Use grant is meant for most research and development work, can cover a whole team, and renews once a year. High-risk Use is a project-level add-on for work blocked under Standard Use; Anthropic says it removes the safeguards that block life-sciences requests, renews every six months, and is available today for Opus 5 and Sonnet 5. High-risk grants for Mythos remain limited to a small set of entities with extra vetting. Cyber classifiers stay in place under the grants.[1]
Two grant types, no BAA org yet
Applicants may seek Standard Use or High-risk Use grants across Claude Science, Claude.ai, Claude Code and the API. The company says the beta is not available for BAA-enabled organizations, so customers with PHI should use separate non-BAA orgs. Newsquawk's headline the same day framed the launch as a biology-related model-access programme, including a beta for teams and institutions. Anthropic opened a beta Life Sciences Verification Program: verified teams get Mythos, Opus and Sonnet with classifiers it says are more permissive for biology work than its generally available Fable models. The company says it has already onboarded dozens of organizations in early access and is opening applications more widely, with individual Pro and Max plans later. A Standard Use grant is meant for most research and development work, can cover a whole team, and renews once a year. High-risk Use is a project-level add-on for work blocked under Standard Use; Anthropic says it removes the safeguards that block life-sciences requests, renews every six months, and is available today for Opus 5 and Sonnet 5. High-risk grants for Mythos remain limited to a small set of entities with extra vetting. Cyber classifiers stay in place under the grants. Newsquawk reported on 17 September 2026 at 17.57 UTC that Anthropic launched a Life Sciences Verification Programme for biology-related model access, including a beta programme for teams and institutions. The page says Standard-use grants apply to Mythos 5.1, Opus 5 and Sonnet 5 models.[1], [2]
Named model versions on the wire
Newsquawk's page says Standard-use grants apply to Mythos 5.1, Opus 5 and Sonnet 5. Anthropic's own post speaks of Mythos, Opus and Sonnet without those point versions in the opening. Newsquawk reported on 17 September 2026 at 17.57 UTC that Anthropic launched a Life Sciences Verification Programme for biology-related model access, including a beta programme for teams and institutions.[2], [1]
