Home › Guides › Open science requirements

Which open science practices are mandatory in Horizon Europe, and which only score points?

Source last checked: 10 September 2026. Quoted from the Horizon Europe Programme Guide V5.1 — 15.09.2025, Part 17 ‘Open science’.

What is mandatory, and what is only recommended?

Horizon Europe draws a hard line that many proposals blur. Some open science practices are mandatory for all beneficiaries per the grant agreement. Everything else is recommended, and recommended practices are incentivised through their evaluation at the proposal stage. The first group is a legal obligation you carry for the life of the project; the second is a scoring opportunity you either take or lose at submission. Part 17, mandatory and recommended practices, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 42

Five obligations sit in the first group. They are set out in the Model Grant Agreement at Article 17, and the Annotated Grant Agreement carries the detailed guidance on complying with them. Part 17, mandatory open science practices, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 43

A work programme or the conditions of a particular call can add to this list. The guide is explicit that some open science practices are mandatory per specific work programmes or call conditions, which may provide for additional obligations. Reading Part 17 is therefore not enough on its own: the call text can raise the floor. Part 17, work programme obligations, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 43

When does each obligation actually bite?

The deadlines are scattered across four pages of the guide and no single table in it collects them. Collected, they look like this.

Two of these catch people out. Publishing is not itself mandatory — if a project intends to exploit its results it may decide not to publish — but the moment a peer-reviewed publication exists, open access is due at publication time, not afterwards. And the DMP is a living document rather than a one-off deliverable: it is expected to be updated after milestones, changes in access provisions, or changes in consortium composition. Part 17, data management plans, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 48

Which licence does each output need?

The licence is not left to the author. The guide names a required minimum set of reuse rights per output type, and the answer differs sharply between text and software.

Part 17, open access and licences, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 52 Part 17, licences for software, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 53

Because the licence obligation binds the author and not the publisher, the guide supplies wording to put in front of a publisher at submission rather than after acceptance. It is worth quoting in full, because it is the one paragraph in Part 17 that a beneficiary can copy and use unchanged.

This work was funded by the Εuropean Union under the Horizon Europe grant [grant number]. As set out in the Grant Agreement, beneficiaries must ensure that at the latest at the time of publication, open access is provided via a trusted repository to the published version or the final peer-reviewed manuscript accepted for publication under the latest available version of the Creative Commons Attribution International Public Licence (CC BY) or a licence with equivalent rights.
Horizon Europe Programme Guide V5.1, Part 17, p. 53

If the publishing agreement turns out to be contrary to the grant agreement obligations, the guide's instruction is blunt: negotiate its terms, and failing that, look for a different publishing venue. Part 17, publishing agreements, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 53

What goes in the proposal, and how long can it be?

Open science is addressed under the Excellence criterion, in the section on methodology, and the guide caps it: information relevant to the specific area of the proposal should be provided in no more than one page. Within that page proposers have to provide concrete information on how they plan to comply with the mandatory practices, and explain which recommended practices they will adopt. Part 17, addressing open science in the proposal, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 44

Two further places in the proposal carry open science weight and are easy to miss. Under ‘Capacity of participants and consortium as a whole’, proposers should show expertise or track record in open science practices relevant to what is planned. And in Part A, proposers list up to five relevant publications, datasets or other achievements — where open access is expected for publications and datasets are expected to be FAIR. Part 17, capacity of participants and Part A, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 44

One line in that passage deserves attention from anyone used to older evaluation habits: the significance of publications will not be evaluated on the basis of the Journal Impact Factor of the venue, but on a qualitative assessment provided by the proposers for each publication. Part 17, assessment of publications, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 44

How is open science scored?

Under two award criteria, not one. Open science practices are evaluated under ‘Excellence’, in particular under methodology, and under ‘Quality and efficiency of implementation’. Part 17, evaluation of open science practices, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 43

The scoring runs in both directions, and the guide states both. Failure to sufficiently address the mandatory practices will result in a lower evaluation score. A clear explanation of how recommended practices will be adopted, as appropriate for the project, will result in a higher one. That is the concrete reason the recommended group is not optional in any practical sense. Part 17, evaluation of open science practices, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 44

Who is outside this?

The ERC. Four separate footnotes in Part 17 carry the same sentence: the information in this section does not apply to the ERC programme, which does not include open science in the evaluation. A fifth adds that the provision on access in cases of public emergency does not apply to the ERC either. Part 17, footnotes on the ERC, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 44

EIC transition calls are inside the policy but scored elsewhere: there, open science practices are exceptionally evaluated under ‘impact’ rather than under excellence. Part 17, footnote on EIC transition calls, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 42

Where this guide stops and the Grant Agreement starts

Part 17 of the Programme Guide is written for the proposal stage. It says so itself: the guide complements the Annotated Grant Agreement with a particular focus on the preparation of proposals, and it defers the operative detail — trusted repositories, the licences required for data, the legitimate reasons for restricting access — to Article 17 of the Annotated Grant Agreement. Part 17, relationship to the Annotated Grant Agreement, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 42

That boundary matters for a second reason. This guide covers what Horizon Europe requires you to open. It does not cover what you may charge for doing it. Cost eligibility, including subcontracting, is a separate obligation in a separate document: it lives in the Annotated Grant Agreement, not the Programme Guide. If your question is about money rather than openness, our guide to Horizon Europe subcontractor costs covers that document instead.

Frequently asked questions

Is a data management plan required with the proposal?

No. The Programme Guide states that a DMP is not required at the proposal submission stage. You describe in summary how you will manage data along the FAIR principles, in the open science section under methodology. The DMP itself is a deliverable of the project and is due by month 6. There is one exception: in cases of a public emergency, and only if the work programme requires it, a full DMP must be submitted with the proposal or at the latest by the signature of the grant agreement.

How long can the open science section of the proposal be?

One page. The guide says information relevant to the specific area of the proposal should be provided in no more than one page, under the Excellence criterion, in the section on methodology. That is the whole budget for describing both mandatory and recommended practices.

What happens if open science does not apply to my project?

You must say so and justify it. The guide is explicit that if proposers believe none of the open science practices apply, they have to provide a justification, and that if evaluators agree, open science will not be taken into consideration in the evaluation. Silence is not the same as justification: failure to sufficiently address open science results in a lower score.

Does open science apply to ERC and EIC proposals?

Not to the ERC. The guide repeats in four separate footnotes that the information in this section does not apply to the ERC programme, which does not include open science in the evaluation, and that the additional provision on access in cases of public emergency does not apply to the ERC either. EIC transition calls are covered but scored differently: there open science practices are exceptionally evaluated under ‘impact’.

Which licence must a Horizon Europe journal article carry?

CC BY. The guide requires open access immediately at publication time under open licences providing specific minimum sets of rights of reuse, and names CC BY for articles and book chapters in edited books. Long-text formats may use CC BY, CC BY-NC, CC BY-ND, CC BY-NC-ND or equivalent. Software is different: with the exception of CC0, Creative Commons licences are not appropriate for software, and the guide recommends licences listed as free by the Free Software Foundation or as open source by the Open Source Initiative.

Is FAIR data the same as open data?

No, and the guide says so directly: FAIR data is not equivalent to open data, and data can and should be FAIR even when access is restricted. FAIR management is required regardless of whether the data will be openly accessible. Open access to data is a separate obligation, governed by the principle ‘as open as possible, as closed as necessary’.

What we do not know

Three things this page deliberately does not tell you.

The Creative Commons rights matrix. Pages 52 and 53 of the Programme Guide carry a checklist showing what a user may do with an output under each CC licence — share, use commercially, adapt, attribute, and what rights are maintained or waived. Our copy of the source captured that table with its rows and columns collapsed together, and we will not reconstruct a rights table by inference. What is stated unambiguously in the surrounding text is the licence-per-output table above. For the matrix itself, read pages 52 to 53 of the guide directly.

What the Annotated Grant Agreement says. Every operative definition in Part 17 — what counts as a trusted repository, which licences data must carry, what a legitimate reason for restricting access is — is deferred by the Programme Guide to Article 17 of the AGA. We have not read that article for this page and do not paraphrase it.

What your call adds. Work programmes and call conditions may impose open science obligations beyond the five above, and some work programmes require deposition in repositories federated in the European Open Science Cloud. No general page can tell you which apply to your call. Part 17, the European Open Science Cloud, Horizon Europe Programme Guide V5.1, 15 September 2025, p. 50

This guide describes Part 17 of the Horizon Europe Programme Guide, version V5.1 dated 15 September 2025, which is the version published by the European Commission and the version checked against the source on 10 September 2026. Part 17 was the part revised in V5.1. The operative obligations themselves sit in Article 17 of the Model Grant Agreement and are detailed in the Annotated Grant Agreement, which this guide does not reproduce.

Written by Robert Grabarević, founder of Scalebiz. He built the radar as a tool for himself, so that he would not have to re-read the same funding rules for every application. Every guide on this site cites the regulation by section or page, so each claim can be checked against the source. About Scalebiz.