Horizon Europe Success Rates 2021-2026: By Pillar, Cluster and Country

Three different numbers get called the “Horizon Europe success rate”, and they disagree by a factor of twenty. The programme-wide average, the figure the press quotes, sits around 14 percent. The average for an instrument such as the ERC or the EIC Accelerator ranges from 6 to 27 percent. And the rate for the one topic you are actually applying to can be anything from 1.7 to 40 percent inside the same call. This page keeps the three apart, gives each with its source, and ends with the arithmetic for the only one that matters to an applicant, the topic-level odds.

What a success rate measures

The Commission counts eligible proposals, not submitted ones, and counts a proposal as successful when it is retained for funding on the main list. Proposals that pass every threshold but fall outside the budget go to a reserve list, and they are the reason the rate is low: around 70 percent of proposals are rated above threshold, while about 16.5 percent are funded, so the constraint is budget and ranking rather than quality. Two consequences follow. A high-scoring rejection is normal, and the resubmission logic in the scoring and ESR guide applies. And the number worth knowing before you write is the ratio of the topic budget to last year’s demand, which the last section shows how to compute.

Programme-wide rate by year

Horizon Europe opened with rates well above its predecessor, then slid as applications rose faster than the budget. The cumulative figure is a reading of the Horizon Dashboard at a point in time, so each row below records when it was read and by whom.

The mechanism is demand. Between the 2024 and 2025 calls, ERC Consolidator applications rose by about 35 percent and Advanced Grant applications by about 31 percent; MSCA Postdoctoral Fellowship applications rose by roughly 65 percent. Budgets did not move with them, so the rates in the next section are lower than the same instruments showed two years earlier, and the 2026-27 work programme, with its budget concentrated on fewer topics, is unlikely to reverse that.

By instrument: the 2025 calls

Pillar I: ERC and MSCA

The ERC keeps rates almost identical across its three domains by allocating panel budgets in proportion to requested budget, so the 12.2 percent Starting Grant figure is also roughly the figure for your panel. What differs is the career-stage trend: Starting 12.2, Consolidator 11.2, Advanced 9.6 percent in 2025, each a point or two below the previous cycle. The eligibility windows and the two-step evaluation that produce these numbers are in the ERC Starting Grant playbook. The MSCA Postdoctoral Fellowship is the sharpest fall in the programme, from 16.4 percent in 2024 to 9.6 percent in 2025 on an application surge; the MSCA fellowship strategy guide covers what that means for the supervisor match and the proposal.

Pillar II: the clusters, call by call

Cluster averages are almost useless for planning because the spread inside one call is enormous. The 2025 Health cluster received 1,444 proposals (749 single-stage and 695 first-stage) for EUR 789 million, and its twelve topics ranged from 1.7 percent to 40 percent. The table lists what the agencies have published for the 2025 calls; where only submissions are known the rate column says so.

Pillar III: the EIC Accelerator

The Accelerator is a funnel. At the 1 October 2025 cut-off, 923 companies submitted full applications requesting almost EUR 7 billion; 121 were invited to interview in January 2026 and 61 were funded, 52 of them with blended finance, for EUR 467 million. That is 6.6 percent end to end but about 50 percent once a company reaches the interview, which is why the short-application and full-application stages, not the pitch, are where most applicants are lost.

Widening and ERA actions

The Widening component has had the programme’s most generous rates, with the exception of the ERA Fellowships once they were flooded with applications.

By country

The Commission’s widening report, with data to 31 March 2025, shows the programme-wide gap between widening and non-widening Member States has nearly closed, at 18.6 against 19.6 percent, with eight widening countries above the average. The gap survives in the two most individual-excellence-driven parts of the programme, the ERC and the EIC, and it is visible in how much excellent work goes unfunded in the south.

For a Greek or Cypriot research office the practical reading is that the collaborative pillars are a level field and the ERC is not, so the consortium route carries better odds than the individual one. The mechanics of that route are in the consortium survival guide, and Proposia’s institutional plans exist because the offices that convert best on this data are in exactly those countries. National funders, agencies and Greece’s own Horizon Europe record are in the guide to research funding in Greece.

Three levers that move your own number

1. Choose topics by budget-to-demand ratio. The 1.7 percent and the 40 percent above sat in the same Health call in the same year. A topic with EUR 80 million and 25 proposals is a different competition from one with EUR 40 million and 118.
2. Use two-stage calls deliberately. Stage 1 passed 24 percent in Cluster 4 and roughly 46 percent of invitees were then funded in Cluster 5, so the compounded odds resemble a single-stage call, but you learn your fate on a 10-page document rather than a 40-page one. The two-stage guide covers the stage-1 format.
3. Look at lump-sum topics. Early Commission assessments reported markedly higher rates for lump-sum topics because risk-averse institutions avoided them; the argument and its source are in Horizon Europe’s hidden instruments.

Estimating the odds for one topic by hand

The Commission publishes, for every topic, the indicative budget and the expected EU contribution per project, and after each deadline it publishes flash call statistics with the number of proposals received. Three lines of arithmetic turn those into a topic-level rate.

The per-project contribution is the number applicants most often get wrong, because the work programme states a range and a typical figure; the budget calculator builds the request bottom-up so you can see which end of the range your consortium actually needs. For live topics with deadlines, budgets and expected contributions in one place, use the open calls database.

Method and refresh schedule

Every row names its source and the date the source was published or read. Official figures from the Commission, the ERC and the executive agencies are used wherever they exist; the two accelopment compilations and the CASRAI dashboard readings are used where the official data is spread across many pages or not published as a table. This page is reviewed quarterly, in January, April, July and October, and the review date at the top changes only when a figure changes. For a cross-funder comparison that includes NIH, NSF and the Nordic funders, see grant success rates by instrument.