Best NIH Grants for Researchers 2026: Odds and Awards

Full answer: The best NIH grant in 2026 is the one your career stage makes you eligible for, because FY2025 odds ranged from 30.5% (K23) down to 10.9% (R21). Graduate students should aim at the F31 ($29,364 stipend, 23.5% success in FY2025). Postdocs should aim at the F32 (26.4%) or, within 4 years of the PhD, the K99/R00 (19.5%). Clinician-scientists should aim at the K08 or K23 (about 30%). Independent investigators are competing for an R01 at 13.0% overall, or 18.9% on resubmission. NIH dropped paylines in January 2026, so a percentile no longer tells you whether you'll be funded.

Last updated: October 7, 2026

Data note: Success rates are from the NIH Data Book success-rate tables for FY2024 and FY2025, the latest published. Award counts and award sizes are from a Grantsights analysis of NIH RePORTER records for FY2024 and FY2025 new awards, pulled on October 7, 2026. Stipends, budget limits and due dates are from NIH notices current on that date.

Best NIH Grants for Researchers in 2026

The Short List: NIH Grants by Career Stage

Seven NIH mechanisms cover most researchers' realistic options in 2026. Grantsights checked each one against its current parent notice, the NIH success-rate tables and the NIH standard due dates on October 7, 2026.

The R01 figures are R01-equivalent rates. The next date shown is for a new application; resubmissions follow a month later (Nov 5, Nov 12 or Nov 16, 2026, depending on the mechanism).

Our view: mechanism choice now matters more than it did two years ago. With R01 odds at 13.0% and each PI capped at 6 applications a year, a postdoc or clinician who's eligible for a K award and skips it to chase an R01 is giving up better than twice the odds.

What Changed at NIH Between 2025 and 2026

Six policy changes since 2025 reshape the decision, and the 2025 version of this guide predated most of them. According to the NIH FY2025 by-the-numbers report, the research project grant success rate fell from 18.5% in FY2024 to 13.0% in FY2025 while applications rose to 62,592.

1. Paylines are gone. In its November 2025 funding-strategy notice, NIH said institutes won't rely on paylines starting with the January 2026 council round. A good percentile is now necessary but no longer a promise.
2. A cap of 6 applications per PI per calendar year. NOT-OD-25-132 took effect with the September 25, 2025 due date. It counts new, renewal, resubmission and revision applications, including multi-PI roles. The same notice says applications substantially developed by AI aren't treated as original work.
3. Multiyear funding. From June 2025, NIH required half of its remaining competing research grant money to go to awards funded for all years up front. NIH's early-stage investigator report cites this as one reason the ESI R01 funding rate dropped from 26.1% to 18.9%.
4. No $500K pre-approval. NOT-OD-26-019 removed both letters of intent and the requirement to get permission before requesting $500,000 or more in direct costs in a year.
5. Simplified review. For due dates on or after January 25, 2025, reviewers score two factors: Importance and Rigor & Feasibility. A third factor, Expertise & Resources, is rated sufficient or not and gets no score. All first-level review moved to the Center for Scientific Review under NOT-OD-25-135.
6. Calendar changes. Separate AIDS due dates ended under NOT-OD-26-029. Fellowships moved to April 8, August 8 and December 8, and NOT-OD-26-064 ended late submissions for fellowships entirely.

One thing that didn't change for applicants: the proposed 15% cap on indirect costs. A Congressional Research Service report from April 2026 says the cap remains blocked in court, a ruling upheld on appeal in January 2026, so negotiated rates still apply.

What the NIH Success-Rate Tables Don't Tell You

The NIH tables show odds, but they hide two shifts that matter for planning in 2026. Grantsights analysis of NIH RePORTER records for new awards (application type 1) shows how sharply the number of new awards fell between FY2024 and FY2025, even where award size held steady.

First, the drop fell hardest on R01s and R21s, and lightest on clinician K awards. New R01s fell by a quarter. K08 and K23 counts fell by 5% to 7%. R00 transitions went up, because those are K99 awardees from earlier years taking faculty jobs. For a clinician deciding between a K23 and a first R01, the data points strongly toward the K23.

Second, award sizes look bigger than they are. The median new R21 award rose from $236,660 in FY2024 to $401,566 in FY2025, and the median direct cost from $150,000 to $272,505. That isn't a bigger R21. It's NIH paying both years of a $275,000 two-year budget at once: 59% of FY2025 R21 awards had a budget period longer than 13 months, against 25% in FY2024. A common mistake is to read FY2025 RePORTER award amounts as annual budgets. Check the project's budget start and end dates before you use a past award to size your own request.

Third, R01 money is concentrated. In FY2025, NHLBI (599), NIAID (460) and NCI (441) administered the most new R01s, and institutions in California, New York and Massachusetts received 32.7% of them. That's a fact about where NIH-funded labs already are, not a rule, but it means applicants at smaller institutions should pay close attention to the Environment part of Factor 3.

The full mechanism-by-mechanism odds by institute are in the Grantsights NIH and NSF funding odds tracker.

1. NIH R01: The Standard Research Project Grant

Best for: Independent investigators with preliminary data and a 4- to 5-year project.

The R01 had a 13.0% success rate in FY2025 across 42,022 R01-equivalent applications, according to NIH Data Book figures. That's down from 18.7% in FY2024 and 21.6% in FY2023. Under the parent announcement PA-25-301, the R01 has no budget cap, runs up to 5 years and uses modular budgets up to $250,000 in direct costs a year.

What a funded R01 looked like in FY2025. The median new R01 carried $449,739 in direct costs and $674,661 in total cost in its first award action, according to Grantsights analysis of RePORTER records. 30.0% of new FY2025 R01s exceeded $500,000 in direct costs, up from 21.0% in FY2024. That jump partly reflects multiyear funding: counting only awards with a normal 12-month budget period, the share was 22.1%.

Resubmission is where R01s get funded. In FY2025, R01-equivalent resubmissions (A1) succeeded 18.9% of the time against 10.8% for first submissions (A0). In FY2024 the split was 27.9% against 15.3%. If your first submission is discussed and scored, the resubmission is usually the better bet than a new idea, and it counts as one of your 6 applications for the year. The NIH resubmission guide covers the one-page introduction.

Budget size now carries institute-level risk. You don't need permission to request $500,000 or more, but institutes still treat large budgets differently. NCI cuts new R01s above $250,000 direct by 17% from the study section's recommended level. NHLBI generally funds R01s for 4 years rather than 5, with exceptions for early-stage investigators and clinical studies. Read your target institute's FY2026 funding policy before you finalize the budget, because underestimating those cuts leaves the aims unfunded in year 3.

Not a good fit: a first-time PI without preliminary data, or a project whose central question could be answered in 2 years for under $275,000. See the NIH R01 guide for the full parent R01 requirements.

2. NIH R21: Exploratory and Developmental Research

Best for: Testing a new direction without the preliminary data an R01 expects.

The R21 allows up to $275,000 in combined direct costs over 2 years, with no more than $200,000 in any single year, under PA-25-304. Preliminary data isn't required. The FY2025 success rate was 10.9%, down from 15.8% in FY2024.

The honest tradeoff. The R21 is no longer the safer route. Its FY2025 success rate (10.9%) was below the R01-equivalent rate (13.0%), for less than half the money and 2 years instead of 5. An R21 makes sense when the idea really is exploratory and an R01 reviewer would mark the absence of preliminary data as a weakness. It's a red flag when a PI uses the R21 as a shorter R01: reviewers notice aims that need 5 years squeezed into 2.

Real FY2025 examples from RePORTER: Tulane University received $219,938 from NIAAA for 1R21AA031728-01. Baylor College of Medicine received $454,791 from NIAID for 1R21AI179603-01A1, a resubmission funded for both years at once. That Baylor award shows why FY2025 R21 dollar totals overstate annual budgets.

R21 and R03 applications are due February 16, June 16 and October 16 for new submissions. For the parent route, see the live NIH parent R21 program page and the NIH R21 guide.

The R03 alternative. The R03 pays up to $50,000 in direct costs a year for 2 years under PA-25-302, and its FY2025 success rate was 17.8%, higher than either the R21 or the R01. The catch: only 14 institutes participate, and NCI joined only in February 2026. See the NIH R03 guide before you build a plan around it.

3. NIH K99/R00: Pathway to Independence

Best for: Postdocs within 4 years of their PhD who are aiming for a tenure-track faculty job.

The K99/R00 succeeded 19.5% of the time in FY2025 (319 awards), down from 25.0% in FY2024 (359 awards). It's a two-phase award under the parent PA-24-194: up to 2 years of mentored K99 support as a postdoc, then up to 3 years of independent R00 support once you're hired as faculty, for no more than 5 years combined.

The numbers people get wrong. The R00 limit is $249,000 a year in total cost, including the new institution's indirect costs, not $249,000 in direct costs. The median award across 353 R00 transitions in FY2025 was exactly $249,000, according to Grantsights analysis of RePORTER records. K99 salary and research support are set by each institute, so there's no single NIH-wide figure. The median new K99 award in FY2025 was $127,345 in total cost.

Eligibility is strict. You need 4 years or less of postdoctoral research experience on the due date. NIH granted two temporary extensions in 2025 and 2026, and both have expired. NCI left the parent K99 in January 2025 and runs its own version under PAR-25-135, which allows up to 6 years of postdoc and pays up to $100,000 a year toward salary plus $30,000 for research.

Real FY2025 examples: the University of Pennsylvania received $173,799 from NIDA for 1K99DA060979-01, and the University of Michigan received $109,255 from NIGMS for 1K99GM157383-01. The $64,544 gap between those two first-year awards is the institute-by-institute variation in action.

New K99 applications are due February 12, June 12 and October 12. The full eligibility rules and review criteria are in the K99/R00 guide.

4. K08 and K23: Mentored Clinician-Scientist Awards

Best for: Physicians and other clinical-doctorate holders building a research career.

K08 and K23 had the best odds of any common individual NIH award in FY2025: 29.7% for the K08 and 30.5% for the K23. The K08 supports mentored laboratory or translational research, and the K23 supports patient-oriented research. Both run up to 5 years and require at least 75% effort (9 person-months) on research, under the parent notices PA-24-182 and its K23 counterpart.

No single dollar figure exists. Each institute sets its own salary and research-support amounts, up to the NIH salary cap of $228,000 for 2026. According to Grantsights analysis of RePORTER records, the median new FY2025 award was $173,572 in total cost for the K08 and $188,307 for the K23. Check your target institute's K award page, not a general summary.

Why the odds are better. The applicant pool is small: 659 K08 and 812 K23 applications were reviewed in FY2025, against 42,022 R01-equivalent applications. Review weighs the candidate, the mentor and the career development plan as heavily as the science. Applications score poorly when the career development plan lists generic coursework instead of naming the specific skill gap each mentor will close.

Pick the branch carefully. Both parents come in clinical-trial-not-allowed and clinical-trial-required versions, and choosing the wrong one gets an application withdrawn before review. Compare the live parent K08 page with the parent K23 page. Institute-level K award odds are in the NIH K award success rates by institute guide, and past K23 winners are listed in the K23 award history.

5. F31: Predoctoral Fellowship

Best for: PhD students who want their own stipend and a fundable record before the postdoc search.

The F31 succeeded 23.5% of the time in FY2025 (849 awards from 3,607 applications), down from 27.9% in FY2024. Under NOT-OD-26-044, the FY2026 predoctoral stipend is $29,364. Tuition and fees are covered at 60% of actual cost up to $16,000 a year, and the institutional allowance is $4,750 at most sponsors. Support runs up to 5 years.

The new calendar catches people. F-series applications are due April 8, August 8 and December 8. The old February, June and October dates are gone, and NIH no longer accepts late fellowship applications for any reason. The next F31 date is December 8, 2026.

The diversity F31 no longer exists. The F31 diversity parent announcement isn't available any more, and the reissued parent F31 has no diversity companion. Older advice to apply through a separate diversity F31 is out of date.

Real FY2025 examples: Mayo Clinic Rochester received $49,538 from NINDS for 1F31NS141217-01. The University of Delaware received $46,329 from NICHD for 1F31HD115311-01A1, a resubmission. Resubmitting is normal for fellowships too.

See the live parent F31 page, and compare institutes in F31 success rates by institute.

6. F32: Postdoctoral Fellowship

Best for: Postdocs early in training who want independent stipend support.

The F32 succeeded 26.4% of the time in FY2025, down from 32.9% in FY2024. The FY2026 stipend runs from $63,480 at 0 years of postdoc experience to $77,076 at 7 or more years, with a $12,400 institutional allowance at most sponsors and tuition up to $4,500. Support runs up to 3 years.

Two institutes are out. NCI and NINDS don't take applications through the parent F32, and NINDS ended its own F32 after June 9, 2026. Cancer and neuroscience postdocs should look at institutional T32 slots or go straight to a K99.

F32 or K99? They aren't either-or, but timing matters. The K99 clock stops at 4 years of postdoc, and a K99 is the stronger signal on the faculty market. Our recommendation: a first-year postdoc with a strong project should file the F32 and plan the K99 for years 2 to 3. A third-year postdoc should put the effort into the K99.

The parent route is the live parent F32 page, and the NIH F31 and F32 guide covers both fellowships side by side.

How Funding Decisions Work Now That Paylines Are Gone

A percentile score was the closest thing NIH applicants had to a funding forecast, and since January 2026 it isn't one. NIH's NCI funding policy page says NCI "no longer uses paylines." The NIAID, NHLBI, NINDS and NIGMS policy pages say the same thing in their own words.

What replaces the cutoff is a funding decision that weighs the review score alongside institute priorities, career stage, the applicant's other NIH funding and portfolio balance. Three practical consequences:

• Talk to the program officer before you submit, not after. Program priority now does work the payline used to do. A short call about fit is worth more in 2026 than it was in 2024.
• Early-stage investigator status still helps at the funding stage. NIAID's FY2026 policy aims for an ESI and new-investigator success rate at or above that of established investigators. ESI status doesn't change your review score.
• Heavily funded PIs face extra scrutiny. NINDS, for example, applies extra stringency to PIs with more than $1.5 million in NIH funding.

Older advice that a 20th percentile gets funded at a particular institute is out of date. For how scores and percentiles are calculated, see NIH grant scores explained. For how institutes set funding before 2026 and what replaced that, see NIH paylines explained.

A 2026-27 Application Calendar

NIH's standard due dates for the next 12 months, from the standard due-dates page and PA-25-301:

A February 2027 R01 submission is reviewed in July 2027, goes to council in October 2027, and can start no earlier than December 2027. Most people underestimate this lag: from first submission to a resubmission award is often close to two years. Plan bridge funding on that basis.

Remember the 6-application limit when you map this calendar. A PI who files an R01, two R21s and their resubmissions has used the year's allotment by summer.

Choosing Your Mechanism: A Decision Checklist

Grantsights compared the FY2025 success rates with each mechanism's eligibility rules to build this order of operations:

1. Are you a trainee? PhD students start with the F31. Postdocs within 4 years of their PhD should plan both the F32 and the K99, and those past 4 years should check NCI's PAR-25-135 or move to faculty-stage routes.
2. Do you hold a clinical doctorate with 75% protected time? The K08 or K23 is the best-odds award NIH offers an individual. Take it before a first R01.
3. Do you have preliminary data for a 4- to 5-year project? Go to the R01, and plan the resubmission from day one.
4. Is the idea genuinely exploratory? Use the R21, accepting odds below the R01. If your institute participates, compare the R03.
5. Count your applications. At 6 per calendar year, every R21 you file is an R01 resubmission you can't.

For a side-by-side of the research grants and K awards, see the NIH R01, R21, R03 and K award comparison. New PIs should also read the NIH new investigator guide. To see open NIH funding opportunities with their current due dates, search the Grantsights grant database.

Sources

• NIH success-rate tables, FY2024 and FY2025: NIH RePORT success rates
• FY2025 research grant totals: NIH Extramural Nexus, FY2025 by the numbers
• Early-stage investigator rates: NIH Extramural Nexus, ESI support in FY2024 and FY2025
• Funding strategy without paylines: NIH unified funding strategy
• FY2026 stipends and allowances: NOT-OD-26-044
• Application limit and AI policy: NOT-OD-25-132
• $500K approval and letters of intent removed: NOT-OD-26-019
• Due dates: NIH standard due dates
• Parent announcements: PA-25-301 (R01), PA-25-304 (R21), PA-25-302 (R03), PA-24-194 (K99/R00), PAR-25-135 (NCI K99/R00), PA-24-182 (K08)
• Award counts, sizes and examples: NIH RePORTER records for FY2024 and FY2025 new awards, queried October 7, 2026