Anthropic Opens Life Sciences Verification in Beta: Looser Bio Safeguards on Opus 5, Sonnet 5 and Mythos 5.1 for Verified Teams
- What happened
- Anthropic launched the Life Sciences Verification Program in beta on 17 September 2026, relaxing biology safeguards for verified teams on the API, Enterprise and Team plans.
- Why it matters
- Standard Use opens Mythos 5.1, Opus 5 and Sonnet 5 (and now Opus 5.5) to routine life-sciences work the classifiers used to block. High-risk Use covers Opus 5 and Sonnet 5; Mythos stays limited to a small vetted set.
- What to do
- Life-sciences teams on a first-party business plan should apply now. Individual plans, third-party platforms and BAA-enabled organisations are not eligible yet.
Our verdict: if your team does drug discovery, research biology or clinical work on Claude through the API, Claude for Enterprise or a Team plan, apply now. Anthropic's Life Sciences Verification Program (LSVP) went into beta on 17 September 2026. It launched as the first public route to loosened biology safeguards on Claude Opus 5, Claude Sonnet 5 and Claude Mythos 5.1, and Anthropic's program page now lists Claude Opus 5.5 too. No directory rating moves: Opus 5 and Sonnet 5 stay recommended, Opus 5.5 and Mythos 5.1 stay conditional.
What happened
Anthropic opened applications for LSVP, initially for teams and institutions. Organisations that pass a review of research credentials, security standards and ethical research oversight get the biology safeguards that block legitimate requests relaxed. Anthropic says it has already onboarded dozens of organisations through early access and expects to enroll hundreds in the first week.
There are two grants:
| Standard Use | High-risk Use | |
|---|---|---|
| Scope | Most biology research: drug discovery, basic science, R&D, supply chain, clinical development, quality assurance, regulatory affairs and investing | Add-on for work that stays blocked under Standard Use; removes all life-sciences safeguards |
| Granted to | The whole team | A single research project |
| Renewal | Once a year | Every six months |
| Models at launch (17 Sep) | Mythos 5.1, Opus 5, Sonnet 5, plus future models as they launch | Opus 5 and Sonnet 5. On Mythos, limited to a small set of entities with additional vetting |
Why it matters
Before LSVP, the biology safeguards blocked some legitimate research requests, and relief was limited to an early-access group. LSVP swaps real-time blocking for offline monitoring of your usage against the use cases you declare. That makes Opus 5 and Sonnet 5 (and, per the program page, Opus 5.5) usable for day-to-day life-sciences work, and opens High-risk Use on Opus 5 and Sonnet 5 for specific projects.
Mythos 5.1 is the exception. Its directory entry already names Life Sciences Verification as one of its two access routes, and that gate is why it is rated conditional. The gate now has a public application form, but for the highest-risk work Mythos stays behind further vetting while Anthropic works with the US government on broader access.
What changes for you
- Eligibility. The first-party console for API usage, Claude for Enterprise and Team plans. Individual plans are not supported yet, LSVP is not yet on third-party platforms, and BAA-enabled organisations are excluded during the beta. Anthropic suggests teams handling PHI use a separate non-BAA organisation.
- Monitoring. Flagged activity is retained for 30 days for review, kept compartmentalised and excluded from model training.
- Other safeguards. Cyber classifiers and every other safeguard stay on.
- Model choice. The announcement says Standard Use extends to future models as they launch, and Anthropic's program page now lists Claude Opus 5.5 and Sonnet 5 alongside Mythos 5.1. It does not say whether High-risk grants cover Opus 5.5, so confirm that with your account team before you move High-risk work off Opus 5.
FAQ
Can I apply on a Pro or Max plan? Not yet. Anthropic says it is working to expand access to individual plans.
Does a High-risk grant cover my whole team? No. It applies to a single research project and must be renewed every six months.
Can I use LSVP through a cloud marketplace? Not yet. It is not available on third-party platforms during the beta.
What to do
- 1 If you are a life-sciences team on the Claude API, Enterprise or Team plan, apply at claude.com/form/life-sciences-verification-program.
- 2 Check your setup against the exclusions first: individual plans, third-party platforms and BAA-enabled organisations are not eligible during the beta. Teams handling PHI should apply from a separate non-BAA organisation.
- 3 Scope any High-risk Use request to a single research project and plan for its six-month renewal.
- 4 Plan for 30-day retention of flagged activity and for offline review of your usage against the use cases you declare.
Affected tools & models
Never need to catch up again
The weekly delta: only verdict changes and act-now items. No digest filler.