Academic publishing operates at a scale measured in hundreds of millions of records and billions of metadata interactions. The figures below describe the infrastructure behind scholarly communication, including Crossref’s records and deposits, open-access availability reported by STM, and selected research-integrity networks. Dates and geography are retained because these statistics describe different populations and measurement periods.
Contents
- The scale of the scholarly record
- Journals, conferences, and publication objects
- Deposits and publishing activity
- Metadata queries and discoverability
- Authors, identifiers, and funders
- Open access in global publishing
- Integrity and shared infrastructure
The scale of the scholarly record
Crossref reported 183,842,464 total records in its database on its Stats Page, updated May 27, 2026. The same report counted 35,006 prefixes and 164,375 journals. A prefix identifies a registrant namespace, while the journal count describes the journal entities represented in the service; neither figure should be read as a count of currently active scholarly titles worldwide.
Crossref’s database also included 140,040 conference proceedings. Together, these figures show why academic publishing statistics need clear definitions: a database record, a journal, a proceeding, and a DOI are different units. Crossref reported 123,859,344 journal DOIs and 9,911,501 conference DOIs on the same date.
The broader record included 9,346,703 components, 507,233 standards titles, and 70,476 databases. Crossref reported 415,700 standards DOIs and 3,475,551 database DOIs. It also reported 201,459 grant DOIs and 1,048,187 peer review DOIs. These categories indicate that scholarly infrastructure covers more than conventional research articles: supporting materials, standards, datasets or databases, grants, and peer-review activity can all have persistent identifiers.
Crossref reported 2,005,894,912 cited-by links. This is a count of links represented in the service, not a count of citations received by every article or a direct measure of research quality. For learners and research managers, that distinction matters when using bibliometric data to compare outputs.
Journals, conferences, and publication objects
Crossref’s title-level DOI counts provide another view of the publishing ecosystem. On the Stats Page updated May 27, 2026, Crossref reported 81,789 journal title-level DOIs and 82,867 conference proceeding title-level DOIs. It also reported 2,436,330 book title-level DOIs, 1,064,995 dissertation title-level DOIs, and 953,987 report title-level DOIs.
Standards and databases were represented at title level as well. Crossref reported 415,135 standards title-level DOIs and 58,383 database title-level DOIs. These counts are not interchangeable with the larger totals for journal DOIs, conference DOIs, or database DOIs. A title-level identifier can describe a collection or continuing publication, while an item-level DOI can identify a particular article, chapter, proceeding contribution, or other object.
| Publication or research object | Crossref count | Measurement period |
|---|---|---|
| Journal title-level DOIs | 81,789 | Stats Page, updated May 27, 2026 |
| Book title-level DOIs | 2,436,330 | Stats Page, updated May 27, 2026 |
| Dissertation title-level DOIs | 1,064,995 | Stats Page, updated May 27, 2026 |
| Report title-level DOIs | 953,987 | Stats Page, updated May 27, 2026 |
| Standards title-level DOIs | 415,135 | Stats Page, updated May 27, 2026 |
| Database title-level DOIs | 58,383 | Stats Page, updated May 27, 2026 |
Crossref also reported 2,947,660 unique DOIs with FundRef data. FundRef data connects DOI records with funding information, making it useful for studying how research outputs relate to sponsors. The figure is a count of unique DOIs with that data, not a count of funders or grants.
Deposits and publishing activity
Deposit activity measures records added to or updated in a registration infrastructure. Crossref’s Stats Page, updated May 27, 2026, reported 12,069,831 added deposits in 2023, 10,886,577 in 2024, and 12,137,579 in 2025. These are annual counts reported by Crossref and should not be treated as a complete census of every scholarly work published globally.
Crossref’s monthly deposit data reported 49,255,880 total records in 2023, 53,544,428 in 2024, and 78,775,016 in 2025. The labels “added deposits” and “total records” describe different measures in the Stats Page, so they should remain separate when assessing activity.
The monthly series for April 2025 through May 2026 shows how activity varied across the reported months:
- April 2025: 999,607 added deposits; May: 980,672; June: 964,716.
- July 2025: 991,670; August: 1,049,095; September: 940,590.
- October 2025: 1,052,464; November: 1,022,102; December: 1,142,863.
- January 2026: 991,539; February: 913,576; March: 1,106,940.
- April 2026: 1,201,468; May 2026: 817,630.
The highest monthly figure in this supplied sequence was 1,201,468 added deposits in April 2026, while the lowest was 817,630 in May 2026. Both figures are Crossref reports from the Stats Page updated May 27, 2026; the latter date is important because it places the May 2026 observation within a particular reporting snapshot.
Metadata queries and discoverability
Metadata services are used not only to register records but also to find and match them. Crossref reported 1,844,795,392 metadata queries in 2023, 1,868,141,440 in 2024, and 2,047,785,984 in 2025. It reported 3,364,991,232 DOI queries in 2023, 3,229,297,408 in 2024, and 3,302,196,736 in 2025.
| Crossref interaction measure | 2023 | 2024 | 2025 |
|---|---|---|---|
| Metadata queries | 1,844,795,392 | 1,868,141,440 | 2,047,785,984 |
| DOI queries | 3,364,991,232 | 3,229,297,408 | 3,302,196,736 |
Crossref separately reported 1,257,433,216 metadata matches in 2025. A query is an interaction with a service, whereas a match indicates that metadata was matched through the reported process. Keeping those measures distinct helps analysts avoid treating every search or lookup as a confirmed identification.
The query figures also show why discoverability is a business-learning issue for publishers, universities, libraries, and research teams. Persistent identifiers and structured metadata support cataloging, reporting, linking, and evaluation. The figures do not reveal the purpose or success of each individual query, so they describe service use at aggregate level rather than reader behavior.
Authors, identifiers, and funders
Crossref reported 40,231,320 total ORCID/article pairs on its Stats Page, updated May 27, 2026. ORCID connects researchers with persistent researcher identifiers, while the article side of the pair connects that identity to a scholarly work. The pair count is therefore an infrastructure measure, not a count of unique researchers or unique articles.
Crossref also reported 2,063,959 total works pushed through ORCID auto-update. The associated permission figures were 5,853,734 authors granting ORCID auto-update permission, 12,386 authors denying permission, and 3,620,282 authors not responding to permission requests. These three response categories describe the reported status of authors in that workflow; they should not be generalized to all academics or all ORCID users.
Funding metadata adds another layer to publication analysis. The 2,947,660 unique DOIs with FundRef data show how persistent identifiers can carry information beyond title, author, and venue. For research administrators, such links can support reporting by sponsor or project, but the supplied figure does not quantify the completeness of funding disclosure across the full scholarly record.
Crossref also reported 112,982,290 journal articles in its database and said members created 6,747,031 journal article records in 2024. Both figures came from Crossref Metadata Awards reporting. The first is a database stock measure, while the second is a creation count for a named year; they should not be combined as though they represented the same period or scope.
Open access in global publishing
STM’s OA Dashboard reported that, in 2024, 40% of all 3.9 million global journal articles, reviews, and conference papers were published as immediately available gold open access. The same dashboard reported that 6% were made freely accessible through green or bronze open access. These percentages refer to the stated global population of 3.9 million journal articles, reviews, and conference papers and to the 2024 measurement period.
STM also reported that gold open access was an option for 80% of global articles in 2024. “An option” describes availability of a publishing route, not the share of articles that actually used that route. Similarly, gold, green, and bronze labels describe access pathways and should not be collapsed into a single rate without preserving their definitions.
The open-access figures are therefore best read as measures of availability and access status within the population STM specified. They do not provide a country-by-country breakdown, a subject-level breakdown, or a forecast. They also should not be applied directly to Crossref’s total record count, because the populations, definitions, and sources differ.
Integrity and shared infrastructure
Research integrity depends on organizations working across publishers, journals, and scholarly communities. STM’s research integrity infrastructure report counted 49 organizational members in the STM Integrity Hub. The same report said that COPE had 106 publisher members representing over 14,500 journals. It also reported 58 organizations in United2Act coordinating responses to paper mills.
Crossref reported that it worked with 124 sponsoring organisations supporting 12,195 sponsored members as of December 2024, and that it added 8 new sponsors in 2024. Its report identified Relawan Jurnal Indonesia, with 3,076 members, and ABEC Brasil, with 1,312 members, among its five largest sponsors. These are program-specific membership figures and do not represent the entire global publisher population.
Taken together, the integrity and sponsorship statistics show a distributed model for scholarly infrastructure. Registration agencies, publisher organizations, journal communities, and research-integrity groups each contribute different services. The counts are most useful when paired with their dates, labels, and scopes: a member count, a journal count, a DOI count, and an organization count answer different questions about how academic publishing functions.