← Evidence & methodology

OpenAI US AISI

1.0.0-rc.1 · Evidence cutoff 12 September 2026

Directly documented interface

Model access

Announced agreement permitting model access; not proof access occurred.

provider → recipient. Provider → recipient; no claim of compulsory access.

Displayed claim → analysis (if present) → coded attributes → interface / instrument → atomic evidence → source version

Instrument and timeline

Instrument
Model-access memorandum
Bindingness
Not established from the announcement
Announced
2024-08-29
Observed by
2024-08-29 · upper-bound
Effective period
UnknownUnknown
Event
Announced
Status at cutoff
unresolved/unknown
Last confirmation
2024-08-29

Later renegotiation announcement is not mapped to a particular 2024 instrument; predecessor continuity unresolved.

Status investigation

Checked 2026-09-12; assessed as of 2026-09-12. Last confirmation does not establish uninterrupted continuity.

amendments
unknown
expiry
unknown
termination
investigated within listed sources
restrictions
investigated within listed sources
succession
see succession.json
conflicts
Later renegotiation announcement is not mapped to a particular 2024 instrument; predecessor continuity unresolved.
Search queries
  • "openai" "us-aisi" access agreement 2026 amendment termination expiry status before:2026-09-13

Analytical assessment

No reviewed analytical conclusion is attached. This does not establish independence or absence of an institutional constraint.

Coded attributes

evaluation-capacity: substantial-access-constrained

Ability to design and execute a substantive assessment independently, conditional on access.

Evaluator × evaluation exercise or programme × date. Claude 3.5 Sonnet 2024 pre-deployment exercise; assessed 2024-11-19.

Completed independent assessment plus explicit access/time limitation satisfies anchor. Not a statement of current CAISI resources.

Supporting evidence: independent-tests access-limits . Counterevidence: independent-tests

  • Coding applies only to stated context and date, not a permanent institutional rank.
Published rubric anchors
absent
Affirmative evidence of no ability to perform this assessment.
nascent
Initial methods/team with no completed substantive assessment.
substantial-access-constrained
Completed independent assessment with documented access or time constraints materially limiting conclusions.
substantial-durable-access
Completed assessment and enforceable/reproducible access sufficient for the task over its defined period.

Unresolved when necessary evidence is missing or adjacent anchors cannot be distinguished. No numeric conversion. Absence requires affirmative remit or operational evidence, not missing sources.

Atomic evidence and exact sources

NIST announced an OpenAI agreement for model access before and after release.

Source
US AISI agreements with Anthropic and OpenAI · NIST
Version
2026-05-04; retrieved 2026-09-12
Exact locator
Opening body, paragraph beginning “Each company’s Memorandum of Understanding”
Limit
Announcement does not establish actual access to every model or current agreement status.

Retrieved bytes SHA-256: 65876c874a485e03147c7467ae419c05dfeded6b31aa7223c6f2c9e83c661b56

The evaluation had a limited access and testing period.

Source
Claude 3.5 Sonnet joint testing report · US AISI / UK AISI
Version
2024-11-19; retrieved 2026-09-12
Exact locator
Section 2.1, printed pages 1–2
Limit
Applies to this exercise; does not identify who caused the limits.

Retrieved bytes SHA-256: cee68f384975897e9ae8fd556baa9ebae9919fe08fe1e1abb363cbead7802903

The institutes each ran independent tests and worked together on methodology and interpretation of findings.

Source
Claude 3.5 Sonnet joint testing report · US AISI / UK AISI
Version
2024-11-19; retrieved 2026-09-12
Exact locator
Section 1, printed page 1
Limit
Independence of tests does not establish durable model access.

Retrieved bytes SHA-256: cee68f384975897e9ae8fd556baa9ebae9919fe08fe1e1abb363cbead7802903

NIST says previously announced partnerships were renegotiated.

Source
New frontier evaluation agreements · NIST
Version
2026-05-05; retrieved 2026-09-12
Exact locator
Second body paragraph
Limit
Bulletin does not identify each predecessor agreement or disclose amended terms.

Retrieved bytes SHA-256: a12961c1edd972ff6d72c45bb8b78793ef9254b57442df684e3fa14a63c77325

Supporting and conflicting evidence

Interface support: openai-access. Counterevidence: caisi-renegotiation

Substantive review

evidence/openai-access: approved · agent reviewer codex-research-checker · 2026-09-12

NIST names OpenAI and a memorandum providing a framework for model access before and after release; access entitlement is not actual access to every model. Checked locator: Opening body, paragraph beginning “Each company’s Memorandum of Understanding”. Source version: 2026-05-04.

Approval fingerprint: 1e7930e6d70cddee066b09a959f5046c89228cfdbf91ecb37db5b43d6212271a

evidence/access-limits: approved · agent reviewer codex-research-checker · 2026-09-12

Full joint report section 2.1 explicitly limits access and the testing window. It does not identify the institution responsible for either constraint. Checked locator: Section 2.1, printed pages 1–2. Source version: 2024-11-19.

Approval fingerprint: 1e7930e6d70cddee066b09a959f5046c89228cfdbf91ecb37db5b43d6212271a

evidence/caisi-renegotiation: approved · agent reviewer codex-research-checker · 2026-09-12

NIST May5 bulletin expressly mentions renegotiated prior partnerships without identifying each antecedent instrument. No automatic continuity or supersession assigned. Checked locator: Second body paragraph. Source version: 2026-05-05.

Approval fingerprint: 1e7930e6d70cddee066b09a959f5046c89228cfdbf91ecb37db5b43d6212271a

instruments/openai-mou: approved · agent reviewer codex-research-checker · 2026-09-12

NIST identifies the OpenAI MoU; underlying legal bindingness and implemented access remain unknown.

Approval fingerprint: 1e7930e6d70cddee066b09a959f5046c89228cfdbf91ecb37db5b43d6212271a

relationships/openai-us-access: approved · agent reviewer codex-research-checker · 2026-09-12

NIST explicitly names the provider-to-US-AISI access framework. Original date retained, actual checkpoint access not inferred, and 2026 unspecified renegotiation does not establish succession or current status.

Approval fingerprint: 1e7930e6d70cddee066b09a959f5046c89228cfdbf91ecb37db5b43d6212271a

institution-attributes/us-sonnet-capacity: approved · agent reviewer codex-research-checker · 2026-09-12

Report methods and constraints support the context-bound evaluation-capacity anchor for the US AISI 2024 exercise. Independent testing counters provider-directed-results inference; no present CAISI resource coding follows.

Approval fingerprint: 1e7930e6d70cddee066b09a959f5046c89228cfdbf91ecb37db5b43d6212271a

evidence/independent-tests: approved · agent reviewer codex-primary · 2026-09-12

Independently inspected the primary source at the recorded locator. Narrow proposition, attribution, date/version and stated limit supported; conflicting access/technical/status accounts remain separate.

Approval fingerprint: 1e7930e6d70cddee066b09a959f5046c89228cfdbf91ecb37db5b43d6212271a

Release-pinned methodology →