The proposed NIH publication cost limit
What NIH floated in July 2025, and how many journals we track would fall under each option.
Checked 23 September 2026
In July 2025 NIH published a request for information asking whether it should cap what a grant may spend on publishing an article, and set out five options — from allowing nothing at all to $6,000 per publication within an award-level cap. None of them has been adopted. An RFI collects comment; it does not make a rule. Of the 753 journals we track with a list price in US dollars, 156 are at or under $2,000 and 746 at or under $6,000.
The five options
| Option as worded in the notice | Per article | Journals we track at or under it |
|---|---|---|
| No publication costs allowed | $0 | 3 journals with a $0 list APC |
| $2,000 per publication | $2,000 | 156 |
| $3,000 if reviewers paid and reviews openApplies only where the journal pays its reviewers and publishes the reviews. We do not verify that for any journal. | $3,000 | 462 if the condition is met |
| 0.8% of direct costs or $20,000 per award, whichever is greater | per award, not per article | not comparable per article |
| $6,000 per publication within award cap | $6,000 | 746 |
Options as stated in NOT-OD-25-138. Counts are our own, over the 753 journals in prices-verified.json whose list price is quoted in US dollars. Thresholds are in US dollars, so only journals whose list price is quoted in US dollars can be compared to them. We do not convert currencies: an exchange rate on the day we publish is not the rate you will be billed at, and a converted figure would look like a verified price when it is not. 646 journals in our data are priced in CHF, EUR or GBP and 153 state no price; all are excluded from every count on this page.
The figures NIH cited
Benchmark figures are NIH’s own, as given in NOT-OD-25-138. We have not recalculated them and they are not drawn from our data.
What it would mean in practice
Most journals we track sit above $2,000
156 of 753 USD-priced journals in our data are at or under $2,000 — 21%. At $3,000 that rises to 462, and at $6,000 to 746. Only 7 sit above $6,000.
A cap is not the same as a bill
A list price is what a publisher quotes before any waiver, institutional agreement or country-tiered adjustment. What a grant is actually charged is often lower. A threshold comparison against the list price is the strict reading, not the likely one.
The $3,000 option has a condition attached
It applies only where the journal pays its reviewers and publishes the reviews. Very few journals do both, and it is not something we verify, so we never mark a journal as meeting that option.
Journals with no author fee are unaffected
A diamond journal produces no publication cost to claim, so no cap can bite. That is the one route that is indifferent to where any of these thresholds lands.