Sources and credits
The public data behind the benchmarks shown on this site, with licenses and requested citations. Only sources that contributed to the published release are listed. Thank you to everyone who builds and maintains them.
Knowledge · Knowledge Depth
Every Knowledge question is built from facts in the sources below, and every fact records where it came from. The counts are the facts each source contributed to the entities of this release (knowledge snapshot kn-v01-2). Questions and answers stay held out; facts from these sources appear publicly only in the development sample, under the license of the source they came from.
- Wikidata4,116 facts
People, places and species: occupations, citizenship, dates of birth and death, coordinates, and taxonomy.
License: CC0 1.0
- ecosyste.ms Packages1,871 facts
Software packages on crates.io, npm, PyPI and CRAN: summaries, first and latest releases, maintainers, licenses and repositories.
License: CC BY-SA 4.0
Package metadata from ecosyste.ms, licensed CC BY-SA 4.0.
- OpenAlex1,861 facts
Research papers: titles, authors in order, publication years, journals, topics and DOIs (snapshot of 2026-09-23).
License: CC0 1.0
Priem, J., Piwowar, H., & Orr, R. (2022). OpenAlex: A fully-open index of scholarly works, authors, venues, institutions, and concepts. arXiv:2205.01833.
- GBIF1,049 facts
Species: scientific names, classification, countries with occurrence records, distribution, habitat and common names.
License: CC0 1.0 or CC BY 4.0, per dataset
Species data from GBIF.org, under the license of each contributing dataset.
- arXiv metadata (via DataCite)1,036 facts
Preprints: arXiv identifiers, titles, authors, first-version dates and categories.
License: CC0 1.0 (metadata)
Thank you to arXiv for use of its open access interoperability.
- ChEMBL965 facts
Medicines: molecule type, development phase, research codes, mechanisms, indications, trade names and approvals (release ChEMBL_37).
License: CC BY-SA 3.0
ChEMBL_37, EMBL-EBI. Zdrazil, B. et al. (2024). The ChEMBL Database in 2023. Nucleic Acids Research 52(D1), D1180–D1192.
- ClinicalTrials.gov809 facts
Medicines: registered trials, sponsors, phases, conditions and start years.
License: Public domain (U.S. National Library of Medicine)
- geoBoundaries (CGAZ)690 facts
Map locations: which administrative units contain a given point.
License: Per country (listed below)
Runfola, D. et al. (2020). geoBoundaries: A global database of political administrative boundaries. PLoS ONE 15(4): e0231866.
- GND, via lobid-gnd543 facts
People: names, dates and places of birth and death, occupations and short biographical notes.
License: CC0 1.0
Gemeinsame Normdatei (GND) of the Deutsche Nationalbibliothek, served by lobid (hbz).
- OpenStreetMap534 facts
Places: what a place lies within, nearby towns, place type, population, elevation and postcode (with Nominatim reverse geocoding).
License: ODbL 1.0
© OpenStreetMap contributors.
- Wikipedia215 facts
Facts from the lead sections of articles, used to build profile fact sheets.
License: CC BY-SA 4.0
Wikipedia contributors.
- Package READMEs122 facts
Software: what a package does, from its own README.
License: Each package's own license
- European Medicines Agency8 facts
Medicines: EU authorisations and indications.
License: Reuse with acknowledgement of the source
Source: European Medicines Agency.
- Drugs@FDA3 facts
Medicines: U.S. approval applications.
License: Public domain (U.S. Food and Drug Administration)
| Source | Used for | Facts | License |
|---|---|---|---|
| Wikidata | People, places and species: occupations, citizenship, dates of birth and death, coordinates, and taxonomy. | 4,116 | CC0 1.0 |
| ecosyste.ms Packages Package metadata from ecosyste.ms, licensed CC BY-SA 4.0. | Software packages on crates.io, npm, PyPI and CRAN: summaries, first and latest releases, maintainers, licenses and repositories. | 1,871 | CC BY-SA 4.0 |
| OpenAlex Priem, J., Piwowar, H., & Orr, R. (2022). OpenAlex: A fully-open index of scholarly works, authors, venues, institutions, and concepts. arXiv:2205.01833. | Research papers: titles, authors in order, publication years, journals, topics and DOIs (snapshot of 2026-09-23). | 1,861 | CC0 1.0 |
| GBIF Species data from GBIF.org, under the license of each contributing dataset. | Species: scientific names, classification, countries with occurrence records, distribution, habitat and common names. | 1,049 | CC0 1.0 or CC BY 4.0, per dataset |
| arXiv metadata (via DataCite) Thank you to arXiv for use of its open access interoperability. | Preprints: arXiv identifiers, titles, authors, first-version dates and categories. | 1,036 | CC0 1.0 (metadata) |
| ChEMBL ChEMBL_37, EMBL-EBI. Zdrazil, B. et al. (2024). The ChEMBL Database in 2023. Nucleic Acids Research 52(D1), D1180–D1192. | Medicines: molecule type, development phase, research codes, mechanisms, indications, trade names and approvals (release ChEMBL_37). | 965 | CC BY-SA 3.0 |
| ClinicalTrials.gov | Medicines: registered trials, sponsors, phases, conditions and start years. | 809 | Public domain (U.S. National Library of Medicine) |
| geoBoundaries (CGAZ) Runfola, D. et al. (2020). geoBoundaries: A global database of political administrative boundaries. PLoS ONE 15(4): e0231866. | Map locations: which administrative units contain a given point. | 690 | Per country (listed below) |
| GND, via lobid-gnd Gemeinsame Normdatei (GND) of the Deutsche Nationalbibliothek, served by lobid (hbz). | People: names, dates and places of birth and death, occupations and short biographical notes. | 543 | CC0 1.0 |
| OpenStreetMap © OpenStreetMap contributors. | Places: what a place lies within, nearby towns, place type, population, elevation and postcode (with Nominatim reverse geocoding). | 534 | ODbL 1.0 |
| Wikipedia Wikipedia contributors. | Facts from the lead sections of articles, used to build profile fact sheets. | 215 | CC BY-SA 4.0 |
| Package READMEs | Software: what a package does, from its own README. | 122 | Each package's own license |
| European Medicines Agency Source: European Medicines Agency. | Medicines: EU authorisations and indications. | 8 | Reuse with acknowledgement of the source |
| Drugs@FDA | Medicines: U.S. approval applications. | 3 | Public domain (U.S. Food and Drug Administration) |
Difficulty tiers use popularity measures from the same sources: Wikidata identifiers and Wikipedia editions (people, species, map locations), population (places), OpenAlex citations (papers, preprints), registry download ranks (software), and registered trials plus PubMed paper counts from the U.S. National Library of Medicine (medicines). No social-media data is used.
geoBoundaries by country
Map-location questions use the geoBoundaries CGAZ boundaries of the 43 countries below. Each country's boundaries come from their own source under their own license.
| Country | Boundary source | License |
|---|---|---|
| AGO | geoBoundaries, INE Angola, citypopulation.de | CC BY 3.0 |
| BDI | RCMRD | Public domain |
| BEL | Namur Capitale, Statistics Belgium | CC BY 4.0 |
| BFA | NYU Spatial Data Repository, Stanford, Map Maker | Public domain |
| BGD | Bangladesh Bureau of Statistics, OCHA ROAP | CC BY 3.0 IGO |
| BIH | Wikimedia Commons | Public domain |
| BOL | GeoBolivia | Public domain |
| CAF | ACTED, OCHA CAR | CC BY 3.0 IGO |
| CAN | Statistics Canada | ODbL |
| CHL | BCN, OCHA ROLAC | CC BY 3.0 IGO |
| CHN | Lee Beryman, OpenStreetMap | ODbL |
| CIV | CNTIG, OCHA ROWCA | CC BY 3.0 IGO |
| CMR | OpenStreetMap, Wambacher | ODbL |
| DEU | Federal Agency for Cartography and Geodesy | Data licence Germany – Attribution 2.0 |
| DZA | OpenStreetMap | CC BY-SA 2.0 |
| ESP | Open Data ESRI, IGN | CC BY 4.0 |
| ETH | Open Africa, Code for Ethiopia | ODbL |
| FIN | OpenStreetMap, Wambacher | ODbL |
| FRA | ARCEP | ODC-By 1.0 |
| GBR | Office for National Statistics | Open Government Licence v3.0 |
| GIN | WFP, OCHA ROWCA | CC BY 3.0 IGO |
| GMB | Knippenberg & Moritz, World Bank Data Blog | CC BY 4.0 |
| GRC | geoBoundaries, Wikimedia Commons | CC0 |
| HTI | FEWS NET | CC BY 4.0 |
| IND | Pathways Data, lgdirectory.gov.in | ODbL |
| ITA | ISTAT | CC BY 3.0 |
| KEN | Benedict Aboki Omare and George Omare | CC BY 4.0 |
| KHM | Open Development Mekong / Cambodia | CC BY 4.0 |
| KOR | Justin Meyers, KoStat | Korea Open Government License Type 1 |
| MMR | Myanmar Analytics Project, USAID | CC BY 4.0 |
| MOZ | INE Mozambique, OCHA ROSEA | CC BY 3.0 IGO |
| MYS | Wikimedia Commons, geoBoundaries | CC BY 4.0 |
| NPL | Open Data Nepal, MoFAGA | CC BY-SA 4.0 |
| NZL | Stats NZ | CC BY 4.0 |
| PAK | Pathways Data, pbs.gov.pk | ODbL |
| PAN | Milton Solano, Smithsonian Tropical Research Institute | CC BY 4.0 |
| PHL | NAMRIA, PSA, OCHA Philippines | CC BY 3.0 IGO |
| PNG | PNG NSO, OCHA ROAP | CC BY 3.0 IGO |
| SEN | Government of Senegal, OCHA ROWCA | CC BY 3.0 IGO |
| SWE | Lantmäteriet, Statistics Sweden | Public domain |
| SYR | UNCS and UNOCHA, OCHA ROSC | CC BY 3.0 IGO |
| TZA | OpenStreetMap | ODbL |
| UGA | UNHCR Operational Data Portal | CC BY 3.0 IGO |
Math · Ballpark
Math questions are generated by code from random seeds, and their answers are computed exactly. No external data is used. See the methodology for how they are scored.