Stripped back, the core questions are always the same: can anyone read this? Who paid to make it available? What can they legally do with it?
What "open access" actually means
Most research is expensive to access. You click a journal link, and a paywall asks for thirty or forty dollars to read a single paper — one you may need for ten minutes, written by a researcher who was paid nothing for it, reviewed by peers who were also paid nothing, and often funded by public money to begin with. Open access is the umbrella term for publishing arrangements that remove that barrier: the paper is online, free, legally readable by anyone with an internet connection.
That sounds simple. Where it gets complicated is in the how. There are multiple routes to open access, multiple business models behind them, and a thicket of terminology that was largely invented by specialists talking to other specialists. The result is that people who would benefit most from understanding this system — students, early-career researchers, journalists, curious members of the public — often bounce off the vocabulary before they reach the substance.
So: here is the whole vocabulary, translated.
| Route | Green | Gold | Diamond |
|---|---|---|---|
| Where the free copy lives | In a repository, deposited by the author | On the journal's own site, from day one | On the journal's own site, from day one |
| Who pays | No one — self-archiving is free | The author, institution or funder (an APC) | A university, society or funder underwrites it |
| The catch | Often the accepted manuscript, sometimes after an embargo | APCs can run from hundreds to many thousands | Usually smaller, more field-specific journals |
The colours (they're routes, not rankings)
Somewhere along the way, the open-access community decided to name its routes after colours. This was not a great communications decision, but the colours have stuck, and they mean real, distinct things.
Green open access is self-archiving. A researcher publishes in a journal — which may or may not be behind a paywall — and separately deposits a copy of their paper in a repository: a publicly accessible digital archive. That repository copy is free to read. The catch is that it is usually not the final published version. Journals typically allow authors to deposit the accepted manuscript — the peer-reviewed text, but before the journal's copyediting and typesetting — rather than the polished PDF you would find on the journal's own site. The difference in content is usually negligible; the difference in formatting can be noticeable. There is also often a delay: the journal may require the author to wait six, twelve, or even twenty-four months before making the deposit publicly visible. This waiting period is called an embargo, and it exists so the journal can sell subscriptions in the meantime.
Repositories where green open access lands include institutional repositories (run by universities and research institutes), and subject repositories covering specific fields. The most prominent subject repositories are arXiv, which hosts preprints and accepted manuscripts in physics, mathematics, computer science and related disciplines; PubMed Central, which does the same for the biomedical and life sciences; and Zenodo, a general-purpose repository run by CERN and the European Commission that will accept research outputs from virtually any field. These are free to deposit in, free to read from, and have been running long enough to be genuinely trusted.
Gold open access is where the published version itself is freely available on the journal's website from the moment of publication — no embargo, no repository hunting required. In principle this sounds ideal. In practice it often comes with a catch of its own: the article processing charge, or APC. Because the journal is no longer earning subscription revenue from readers, it recoups costs by charging authors (or more precisely, their institutions or funders) a fee to publish. APCs vary enormously, from a few hundred dollars at the lower end to well over ten thousand at some high-prestige journals. For a researcher without institutional support or grant funding to cover that fee, gold open access can mean paying a substantial sum out of pocket — or not publishing open access at all. This financial structure has real consequences for who gets heard in global scholarship.
Diamond open access is gold without the APC: the published version is freely available, and nobody pays a processing charge. Diamond journals are typically funded by universities, learned societies, library consortia, or national research bodies. They are genuinely free on both sides — free to read, free to publish in. They tend to be smaller and more field-specific than the big commercial journals, but their number is growing, and the model is increasingly seen as the most equitable form of open access. The Directory of Open Access Journals, known as DOAJ, indexes thousands of peer-reviewed open-access journals and is a useful starting point for finding reputable diamond and gold titles in any discipline.
A fourth term you may encounter is hybrid open access: a subscription journal that offers authors the option of paying an APC to make their individual paper freely available within an otherwise paywalled journal. Critics call it double-dipping, because the institution pays a subscription and an APC for the same journal. It is widespread and controversial, and most open-access advocates regard it as a transitional arrangement at best.
| Licence | Attribution | Commercial use | Modifications |
|---|---|---|---|
| CC BY | Required | Allowed | Allowed |
| CC BY-SA | Required | Allowed | Allowed, under the same licence |
| CC BY-NC | Required | Not allowed | Allowed |
| CC BY-ND | Required | Allowed | Not allowed |
Licences: what "free to read" actually permits
Free to read does not automatically mean free to reuse. This is where Creative Commons licences come in, and understanding them matters if you want to quote from, reproduce, translate, or build on open-access work.
Creative Commons is a non-profit that created a suite of standardised, human-readable licences. The most permissive is CC BY (Attribution): you can do almost anything with the work — copy it, adapt it, use it commercially — as long as you credit the original author. This is the licence most funders and open-access advocates prefer, because it maximises downstream use. More restrictive variants add conditions: CC BY-SA requires derivative works to carry the same licence (ShareAlike); CC BY-NC prohibits commercial use (NonCommercial); CC BY-ND prohibits modifications (NoDerivatives). A paper marked CC BY-ND is open to read but not to adapt — an important distinction if you are, say, translating research for a public audience or creating teaching materials from it.
When no licence is stated at all, default copyright applies: the paper may be free to read (if the journal made it so), but legally you cannot reproduce it beyond what fair use or fair dealing permits in your jurisdiction. Always check the licence, not just whether you can see the paper.
- Submitted manuscriptThe author's first version, sent to a journal for consideration.
- Under peer reviewIndependent experts assess the work and ask for revisions.
- Accepted manuscriptThe peer-reviewed text — what green open access usually deposits.
- Version of recordThe final, typeset paper the publisher produces and hosts.
- Repository depositA free copy lodged in an archive, with a persistent identifier.
Preprints: before peer review, still useful
A preprint is a version of a paper that has been shared publicly before it has been through formal peer review. Authors post preprints to repositories like arXiv or Zenodo to share their findings quickly — useful in fast-moving fields — while the peer-review process grinds on in parallel, often for months. Preprints are not the same as published, peer-reviewed papers, and they should be treated accordingly: as works in progress that may change, be corrected, or occasionally be retracted. They became widely visible to general audiences during the pandemic, when Covid-19 research was moving too fast for traditional publication timelines, with mixed results.
The distinction between a preprint, an accepted manuscript, and a version of record (the final, typeset, journal-published paper) matters when you are citing, sharing, or evaluating research.
Mandates, embargoes and the compliance maze
A mandate is a policy that requires open access. Funders — research councils, government agencies, charitable foundations — increasingly mandate that research they fund must be made openly available, either immediately or within a set period. Institutions have their own mandates too. The details vary: some require gold or hybrid open access with a CC BY licence; others accept green open access via a repository; some specify particular repositories. Researchers navigating these requirements can find themselves consulting multiple policies simultaneously.
An embargo, as noted above, is the period during which a publisher prevents or delays the deposit of an accepted manuscript in a repository, even if the author would otherwise be free to do so. Embargo lengths have been a battleground between publishers and open-access advocates for years. Some funders now refuse to fund embargoes at all, requiring immediate open access as a condition of their grants.
Transformative agreements are a more recent development: deals between institutions (usually via library consortia) and publishers that aim to shift subscription spending toward open-access publishing. Under a transformative agreement, researchers at a subscribing institution can publish open access in the publisher's journals without paying individual APCs, because the institution's deal covers it. They are common in Europe and increasingly elsewhere, though their long-term sustainability is debated.
Where all of this leaves you
The open-access landscape is genuinely complicated, and the jargon has not helped. But stripped back, the core questions are always the same: Can anyone read this? Who paid to make it available? What can they legally do with it?
Green, gold, and diamond describe who bears the cost and where the free version lives. APCs and embargoes describe the friction points. Licences describe what freedom you actually have once you get to the paper. Preprints are the fast lane with caveats attached.