EB-1A requirements in 2026: the 10 criteria, what counts as evidence, and the test nobody posts about
You have seen the list before — "any 3 of 10." What the forums rarely explain is that in 2026 meeting three criteria is the entry ticket, and with EB-1A approvals at a record-low 32.2% (verified quarter data), the fight is won or lost at the final merits stage. This page covers each criterion with the 2026 evidence bar, then the merits test itself.
First: the two-step test
Step 1 — criteria count. At least 3 of the 10 regulatory criteria (8 CFR 204.5(h)(3)), or one one-time major achievement like an Olympic medal or Pulitzer.
Step 2 — final merits. The officer decides whether those evidences, together, prove you are one of the small percentage at the top of your field, with sustained acclaim, who is coming to continue work in the US. Ten thin meets-items can lose here; four deep ones can win. This is where most 2026 denials happen.
The 10 criteria and the 2026 evidence bar
1. Prizes or awards (national/international)
What counts: competitive awards with published criteria and a named jury. Independent of employer internals.
2026 reality: "best paper at a regional workshop" and company service awards get waved through the count but weigh near-zero at merits. Documentation must show selectivity: number of entrants, criteria, judging panel.
2. Membership requiring outstanding achievement
What counts: memberships where admission itself requires excellence — IEEE Senior Member, national academies, jury-reviewed artist guilds. The criteria for admission must be published and you must show you met them.
2026 reality: pay-the-dues memberships (ACM, most "professional" associations) do not qualify at all. Officers check admission criteria first.
3. Published material about you in the press
What counts: articles, interviews, profiles where you or your work are the subject, with byline, date, and circulation evidence. Trade press counts if it has editorial standards.
2026 reality: the most over-claimed criterion. A quote inside someone else's article, a company newsletter, a PR wire release or a Medium post is not "published material about you." Sponsored/contributed content is actively harmful — it reads as self-promotion.
4. Judging the work of others
What counts: peer review for journals/conferences, grant panels, competition judging, thesis committees outside your own institution. Letters from editors confirming review volume work.
2026 reality: one-off internal reviews don't carry. Two-plus documented review streams per year, verifiable by third parties, do. Easiest criterion to build in 3–6 months if missing.
5. Original contributions of major significance
What counts: contributions adopted by others: licensed, used in standards, productized at scale, cited by independent leaders with explanatory context.
2026 reality: the hardest criterion, and the one "major significance" kills. You must prove the impact of the work, not the existence of the work. Self-citations and colleague-citation rings are increasingly spotted.
6. Authorship of scholarly articles
What counts: peer-reviewed journal/conference papers, scholarly books, articles in noteworthy professional media. Citation record with independent context letters helps.
2026 reality: raw count matters less than where and who cites. Predatory-journal papers are now actively discounted — check your venues.
7. Display of your work (arts)
What counts: exhibitions at reputable galleries, festivals, showcases, published architecture, catalogued distribution.
2026 reality: open-submission and pay-to-show venues carry almost nothing. Juried venues with acceptance rates documented carry a lot.
8. Leading or critical role for distinguished organizations
What counts: director/VP/founding-team roles where the org is distinguished (not just big) and your role is documented as essential to outcomes, with org-level proof of distinction.
2026 reality: "lead engineer" without org distinction evidence is the classic miss. The org's reputation must be proven in its own right.
9. High salary compared to peers
What counts: top-percentile compensation documented via government labor statistics, industry commission data, or named reliable surveys — plus context of role/level.
2026 reality: Glassdoor screenshots are weak. Total comp vs base comp games get read. For athletes/artists: contracts, royalties, ticket/chart data.
10. Commercial success in the performing arts
What counts: box office, sales, streaming metrics, record charts with third-party verification.
2026 reality: rarely claimed outside entertainment; when claimed, unsourced numbers are the RFE trigger.
The pattern in every 2026 denial
- Laundry lists: many claims, thin proof each — fails final merits even with 3+ "met."
- Circular letters: recommenders who are co-authors, colleagues, or the applicant's own company.
- Self-published "press" and contributed advertorials.
- Claiming criteria that don't fit the field's evidence culture (salary claims on non-market careers).
See which criteria your record actually meets — free evaluator
Educational summary of 8 CFR 204.5(h)(3) and published adjudication trends — not legal advice, not a prediction of any individual case.