Anthropic researcher access programs (web)

Citation

Anthropic. “What is the External Researcher Access Program?” Claude Help Center article, dated March 16, 2026. https://support.claude.com/en/articles/9125743-what-is-the-external-researcher-access-program.

Anthropic, PBC. “Anthropic’s AI for Science Program Official Rules.” Last updated Aug. 25, 2026. https://www.anthropic.com/ai-for-science-program-rules.

Both fetched 2026-10-15 by the agent. Read 2026-10-15.

Bottom line

Anthropic has two credit programs. The External Researcher Access Program gives $1,000 of free API (application programming interface) credit for AI safety and alignment research. The AI for Science Program gives credit for science research, in an amount that Anthropic sets.
The External Researcher Access Program reviews on “the first Monday of each month”. It does not reply to unapproved applicants.
Both programs keep the usage rules in force. The AI for Science rules add “a biosecurity assessment” of each application.

Background

The External Researcher Access Program article is in the Claude Help Center, under the API FAQ. It was “substantially revised in July 2024”. The AI for Science rules are the legal terms of a separate program. The rules do not describe the application form. A Help Center article on the AI for Science Program exists.

Key claims

External Researcher Access Program (Help Center article):

Purpose: the program supports “researchers working on AI safety and alignment topics that we consider high priority by providing free API credits.” (Introduction)
Application: “Please complete the following application form with details about your team and research topic”. (Introduction)
Scope: the credits are “for API use and do not apply to the Claude web app.” The program “does not provide access to non-standard or nonpublic models.” (About this Program, items 2 and 3)
Use rules: “Applicants through this program do not receive exemption from our Usage Policy.” (About this Program, item 4)
Review cycle: “We evaluate submissions on the first Monday of each month.” (About our Process, item 1)
Size: “If successful, we will apply $1000 in API credits to the account. For rare special cases, we may approve a higher quantity of credit.” (About our Process, item 2)
No reply on decline: “we regret that we cannot provide individual responses to unapproved submissions.” (About our Process, item 3)
Volume: “We receive thousands of applications each month.” (FAQ, “I have contacted you…”)
Paid tier: “it does not block them from simply purchasing API credits in the interim.” (About our Process, item 1)
Group accounts: “we’ll allocate credits to a single Claude Console organization ID, which you can then flexibly distribute among your researchers.” (FAQ, “I run a research program…”)
Other fields: “we’re unable to extend free API credits to other fields of research at this time”. Researchers in other fields can “purchase API credits”. (FAQ, “Does your Researcher Access Program…”)
Exemption path: a close collaborator of an Anthropic employee can ask for a Usage Policy exemption through an “internal approval process”. (FAQ, red-teaming question)

AI for Science Program (Official Rules):

Discretion: “Anthropic will choose the winning researcher in its sole discretion”. (Section 1)
Eligibility: an individual “must be at least 18 years of age”. An employee of an institution warrants “that you have obtained all consents from that employer or institution necessary to participate”. The rules exclude residents of listed sanctioned places. (Section 2)
Selection: “Each Selected Researcher will be selected based on an objective evaluation of their application responses.” (Section 4.1)
Biosecurity: “Applications will also undergo a biosecurity assessment to ensure they do not raise concerns around accelerating harmful applications.” (Section 4.2)
Ranking: “the highest-scoring eligible applications receiving Credits.” (Section 4.2)
Size: “Selected Researchers will receive credits applicable to the relevant Anthropic offering, as determined by Anthropic in its sole discretion”. (Section 6)
Duties: “Selected Researchers will participate in research surveys and interviews conducted using Anthropic’s tools”. (Section 8)
Publicity: researchers grant permission to use their “names, biographical information, photographs, voices and/or likenesses for advertising and promotional purposes”. (Section 7)
Change: “Anthropic reserves the right in its sole discretion to cancel, terminate, modify, or suspend the Program.” (Section 10)
Law: “These Rules and the Program shall be governed by the laws of the State of California”. (Section 13)

Evidence and method

Both are Anthropic’s own pages. The Help Center article states the process. The rules state the legal terms. Neither page gives the scoring criteria in detail.

Limits

Neither page states who reviews. The rules say only “Anthropic”.
The External Researcher Access Program article does not state the duration of the credit or a renewal path. It invites an “updated application” when needs change.
The AI for Science rules do not state a credit amount, a duration or the application steps.
Neither page states an acknowledgement line for publications.
Neither page states eligibility by institution type.

Source file: wiki/sources/Anthropic researcher access programs (web).md, updated 2026-10-15.