evaluator; distance from a lab: 1
SecureBio
Scorecard
Main biosecurity evaluator; author of the Virology Capabilities Test; founding member of the AI Evaluator Forum.
Nonprofit, Cambridge, US. Confidence high. Domains: bio, misuse. Scores by preset: Lab procurement 71; Regulator or auditor 71; Public trust 69; Equal weights 72. Weakest dimension: funding 2/4.
What would move the score. Publish the AI division's client mix and the written terms that insulate it from Detection's funders.
Funding 2/4
Labs pay per engagement; otherwise diversified.
evidence: tier 3 (self)
Anchor 2 on funding given the cited signals.
- Labs typically pay for evaluations of their own models; OpenAI covered the cost of the GPT-5.5 assessment.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed2026-09-14 - Accepted an OpenAI Foundation grant in 2026 for the Detection division; the foundation's endowment is a stake in OpenAI and its board largely overlaps.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed2026-09-14 - The OpenAI Foundation granted SecureBio Detection $17.2M; the Foundation's endowment is a stake of roughly a quarter of OpenAI PBC with a near-identical board.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed SecureBio on X: OpenAI Foundation grant (SecureBio, 2026) tier 3 selfconfirmed SecureBio on X: OpenAI Foundation grant (SecureBio, 2026) tier 3 selfconfirmed2026-09-15 - Coefficient multi-year grant and AI Safety Fund grants; US CAISI Bio R&D contract and EU AI Office contract in 2025.
SecureBio AI: 2025 in Review (SecureBio, 2026-01) tier 3 selfconfirmed2026-01 - Open: Coefficient grant amount to SecureBio.
Governance 3/4
Nonprofit or public body with a COI policy.
evidence: tier 3 (self)
Anchor 3 on governance given the cited signals.
- Two-division structure with a stated firewall; leadership committed publicly to resign and disclose if the grant were used as leverage on evaluations.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed2026-09-14 - Founding member of the AI Evaluator Forum, which wrote the AEF-1 independence standard.
AI Evaluator Forum launch and AEF-1 (AI Evaluator Forum, 2025-12-04) tier 3 selfconfirmed2026-09-14 - The grant is restricted to Detection work; the AI evaluation team has separate leadership, budgets and deliverables.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed Building a three-day early-warning system for novel pathogens (SecureBio (Substack), 2026) tier 3 selfconfirmed2026-09-15
Personnel 3/4
Recusal policy and disclosure of lab ties.
evidence: tier 4 (press)
Anchor 3 on personnel given the cited signals.
- Staff pipeline disclosed through philanthropic career grants; no lab board roles found.
It looks like there are some good funding opportunities in AI safety right now (80,000 Hours, 2025-10-13) tier 4 pressconfirmed2026-09-14
Access depth (lab-granted) 3/4
Helpful-only or weights-level access, chain of thought, logs, on-site.
evidence: tier 3 (self)
Anchor 3 on access given the cited signals.
- Pre-deployment access across the major Western labs.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed2026-09-14
Scope control 3/4
Evaluator sets scope and can add questions.
evidence: tier 3 (self)
Anchor 3 on scope given the cited signals.
- Evaluates models from many developers; consortium member on the EU AI Office CBRN contract.
FAR AI (Longterm Wiki, 2026-02-26) tier 4 pressconfirmed2026-09-14 - SecureBio states the OpenAI Foundation grant places no constraints on what any part of SecureBio can research, publish or say, and its AI team's policy is to evaluate models solely on their merits.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed2026-08-20
Publication rights 3/4
Publishes; redaction limited to security; redactions disclosed.
evidence: tier 3 (self)
Anchor 3 on publication given the cited signals.
- Publishes pre-release biological assessments and the Virology Capabilities Test.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed AI Evaluator Forum launch and AEF-1 (AI Evaluator Forum, 2025-12-04) tier 3 selfconfirmed2026-09-14
Method transparency 3/4
Tasks or code partly open.
evidence: tier 3 (self)
Anchor 3 on methods given the cited signals.
- Methods published; VCT is documented.
AI Evaluator Forum launch and AEF-1 (AI Evaluator Forum, 2025-12-04) tier 3 selfconfirmed2026-09-14
Role incompatibility 3/4
Tools are open or free to the ecosystem.
evidence: tier 3 (self)
Anchor 3 on products given the cited signals.
- No commercial products found; tools shared with governments.
Building a three-day early-warning system for novel pathogens (SecureBio, 2026-08-20) tier 3 selfconfirmed2026-09-14
Traced money and ties
hop 0: 1 row; hop 1: 1 row (grant $17.2M); hop 2: 1 row (recommendation $0.8M). Confirmed rows: 3 of 3. Dollar sums: confirmed + unaudited USD rows only; imported figures are not re-derived and never summed. Second hop traced for 3 of 3 sources. Ties: none within two steps recorded. Bounded negatives on file: 0.
Ledger
Money in
- T22 from OpenAI Foundation: grant $17.20M, 2026. grant to SecureBio Detection, restricted to Detection work; $17.2M confirmedtier 3 selfhttps://securebio.org/blog/three-day-early-warning-system/
- T23 from OpenAI: contract undisclosed, 2026. OpenAI covered the cost of the GPT-5.5 evaluation confirmedtier 3 selfhttps://forum.effectivealtruism.org/posts/dMfgJQ2rGX8GmzdMm/thoughts-on-taking-openai-foundation-funding
- T69 from Survival and Flourishing Fund: recommendation $0.75M, 2025. SFF 2025: $754,000 general support; $1,500,000 for SecureDNA lists SecureBio as receiving charity (not added, different program) confirmedtier 3 selfhttps://survivalandflourishing.fund/2025/recommendations
One hop away
Funding graph, focused
Neighbours at full strength, everything else faded. Hover any name to move the focus; click a name to open its page.