Kinds of data

Text and archives

Published and unpublished writing: books, reporting, journals, specialist publications, community archives.

Public price: $881 to $3,082
per book or title
Demand: Steady
Who buys: Labs, answer engines, national language programmes

The market rate for text, by side

The median where three or more figures exist, otherwise the figures we hold. Each side is a different point in the market: what the people who make it are paid, what holders are offered, what catalogues ask, what was paid.

Paid to the people who make it

$52

per hour of work, median of 3 figures; middle half $40 to $70

Offered to holders

$355

per book or title, median of 3 figures; middle half $280 to $1,428

Sales that happened

$2,750

per book or title, median of 10 figures; middle half $881 to $3,082

Asking prices off the shelf

$0.001

per word, median of 9 figures; middle half $0.00011 to $0.0025

Asking prices off the shelf

$0.013

per record, median of 6 figures; middle half $0.0048 to $0.018

Asking prices off the shelf

$0.22

per article, median of 5 figures; middle half $0.21 to $0.22

Asking prices off the shelf

$78

per book or title, median of 4 figures; middle half $60 to $80

Asking prices off the shelf

$4.88 to $49

per document or page, from 2 figures

Asking prices off the shelf

$0.00000082 to $0.0000011

per token, from 2 figures

Made to order

$0.02

per record, from 1 figure

Made to order

$129

per book or title, from 1 figure

Cost to produce

$0.0000024

per token, from 1 figure

What text has sold or been offered for

The whole price board

Public prices for text
What was sold or offered	Price	When	What to know	Source
Silicon-DPO Platinum, The Reasoning-Code Dataset: 5,001 preference pairs with think-trace and self-corrected code, verified by Python unit tests in a sandbox (coding 35%, finance 20%, STEM 15%; JSONL), Commercial / Startup licence, sold on Gumroad	€499 per dataset of 5,001 pairs (one-time)	Oct 2026	The commercial tier of a two-tier listing by one individual seller; undated page, read 2026-10-17. Only 35% of the pairs are coding, so it is filed as a text preference dataset, not a code dataset, although the listing calls it a reasoning-code dataset. Gumroad's own page shows a USD price of 55.08 for the EUR 49 tier (1.124 USD per EUR); the ECB rate is used here. About ten times the research tier.	Gumroad, Silicon-DPO Platinum: The Reasoning-Code Dataset (storefront ryanride) A catalogue or marketplace listing
Silicon-DPO Platinum, The Reasoning-Code Dataset: 5,001 preference pairs with think-trace and self-corrected code, verified by Python unit tests in a sandbox (coding 35%, finance 20%, STEM 15%; JSONL), Personal / Research licence, sold on Gumroad	€49 per dataset of 5,001 pairs (one-time)	Oct 2026	The non-commercial tier of the same listing, kept apart from commercial prices. Undated page, read 2026-10-17.	Gumroad, Silicon-DPO Platinum: The Reasoning-Code Dataset (storefront ryanride) A catalogue or marketplace listing
Humanities expert work on Mercor (standing talent network; band its listings have posted at)	$45 to $60 an hour per hour	Oct 2026	The posting closed; its page now shows the standing Mercor talent network for the field, which states the band its listings have paid and names a project's own rate only when it invites someone (was $50 an hour on the posting).	Mercor, English Language and Literature Expert (closed) and its Humanities Expert talent network The payer's own page
NLP entity annotation (named entity labelling), Label Your Data starting price	$0.02 per entity for NLP tasks per entity, starting price	Oct 2026	A floor ('starts as low as'), not a typical price; undated page read on 2026-10-14.	Label Your Data, Pricing (FAQ) A vendor's published price guide
AI language expert, French (native speakers with teaching experience) on Alignerr	$25-30/hr per hour	Oct 2026	The language tier sits well below expert and coding tiers on the same board ($40 to $120 an hour).	Alignerr, open roles The payer's own page
100,000+ rights-cleared academic and professional e-books (PDF, EPUB; English, Hindi and other languages) from Arts & Science Academic Publications	$53,514,000 per collection of 100,000+ e-books (non-exclusive, perpetual)	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. A publisher selling its own catalogue; listed 24 Jul 2026.	Opendatabay, 100K+ EBooks related to STEM & Non-STEM in multiple languages A catalogue or marketplace listing
18,000+ rights-cleared STEM and non-STEM EPUB e-books from Indian publishers, licensed by Arts & Science Academic Publications	$9,473,700 per collection of 18,000+ e-books (non-exclusive, perpetual)	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 24 Jul 2026.	Opendatabay, 18000+ EBooks in epub format STEM & Non-STEM A catalogue or marketplace listing
Quora questions and answers dataset (10.2M+ records, 21 fields) scraped and sold by Bright Data	Up to $0.0025 per record, Min Order $250 per record	Oct 2026	Scraped forum text; the buyer carries the rights question, which is why it prices like a collection fee rather than a licence.	Bright Data, Quora datasets A catalogue or marketplace listing
Wikipedia articles dataset (9.2M+ records, 11 fields) scraped and sold by Bright Data	$250 for 100K records (approximately $0.0025 per record) per 100,000 records (one-time)	Oct 2026	Entry price for a one-time snapshot of openly licensed text; the price is for collection and delivery, not rights.	Bright Data, Wikipedia datasets A catalogue or marketplace listing
60M+ words of editorially curated magazine text (about 90 issues, English, Hindi and Indian regional languages), 100% cleared IP	$150,000 per corpus of 60M+ words	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. A single publisher's own archive (Prakhar Goonj Publications), proprietary licence. Listed 2 Oct 2026.	Opendatabay, 60M+ Words Multilingual AI/LLM Text Corpora (100% Cleared IP) A catalogue or marketplace listing
Al-Hayat Arabic newspaper corpus (42,591 articles in 7 domains, 18.6M tokens)	1440.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, Al-Hayat Arabic newspaper corpus (ELRA-W0030) A catalogue or marketplace listing
Amharic-English bilingual news and legal corpus (232,653 Amharic and 291,701 English words)	4000.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence. Per word counts both sides (232,653 + 291,701). A low-resource language pair prices about 100 times the per-word rate of big newswire sets.	ELRA catalogue, Amharic-English bilingual news and legal corpus (ELRA-W0074) A catalogue or marketplace listing
An-Nahar Arabic newspaper text, 1995 to 2000 (6 years, 45,000 articles and 24 million words a year)	4788.00 € per corpus, research use only	Oct 2026	ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence. Word count is 6 years x 24 million words as stated; one year alone is 1,008 EUR for a commercial organisation.	ELRA catalogue, An-Nahar Arabic newspaper text, 1995 to 2000 (ELRA-W0027) A catalogue or marketplace listing
Annotated tweet corpus in Arabizi, French and English (585,163 tweets, opinion and theme annotations)	10000.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, Annotated tweet corpus in Arabizi, French and English (ELRA-W0323) A catalogue or marketplace listing
Chinese-Vietnamese Parallel Corpus (200,000 sentence pairs, average 15 words per sentence)	2100.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, Chinese-Vietnamese Parallel Corpus (ELRA-W0312) A catalogue or marketplace listing
CRATER corpus (aligned telecom manuals, 1,000,000 tokens per language in English, French and Spanish)	100.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence. Per word counts 3 languages x 1,000,000 tokens.	ELRA catalogue, CRATER corpus (ELRA-W0003) A catalogue or marketplace listing
Dutch PAROLE Distributable Corpus (3 million words of books, newspapers and periodicals)	2500.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, Dutch PAROLE Distributable Corpus (ELRA-W0019) A catalogue or marketplace listing
ECI/MCI European Corpus Initiative Multilingual Corpus I (over 98 million words in most major European languages)	50.00 € per corpus, research use only	Oct 2026	ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence. A distribution fee rather than a value price.	ELRA catalogue, ECI/MCI European Corpus Initiative Multilingual Corpus I (ELRA-W0004) A catalogue or marketplace listing
Archives of El Mundo newspaper, 2020 to 2022 (45,658 Spanish articles, El Mundo and sister titles)	8850.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence. The rare case of a recent newspaper archive with a public commercial price; members pay 5,900 EUR.	ELRA catalogue, Archives of El Mundo newspaper, 2020 to 2022 (ELRA-W0332) A catalogue or marketplace listing
Archives of El Mundo newspaper, year 2020 (15,073 Spanish articles)	3000.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, Archives of El Mundo newspaper, year 2020 (ELRA-W0333) A catalogue or marketplace listing
Archives of El Mundo newspaper, year 2021 (14,461 Spanish articles)	3000.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, Archives of El Mundo newspaper, year 2021 (ELRA-W0334) A catalogue or marketplace listing
Archives of El Mundo newspaper, year 2022 (16,124 Spanish articles)	3000.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, Archives of El Mundo newspaper, year 2022 (ELRA-W0335) A catalogue or marketplace listing
English-Vietnamese Parallel Corpus (1,000,000 sentence pairs, average 20 words per sentence)	12000.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence. About $0.0003 per word if both sides average 20 words.	ELRA catalogue, English-Vietnamese Parallel Corpus (ELRA-W0311) A catalogue or marketplace listing
Monolingual Greek corpus (1 million words of 1996 Eleftherotipia newspaper articles)	600.00 € per corpus, research use only	Oct 2026	ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence.	ELRA catalogue, Monolingual Greek corpus (ELRA-W0014) A catalogue or marketplace listing
ICE-GB, British component of the International Corpus of English (a million words of spoken and written English, fully parsed)	1500.00 € per corpus, research use only	Oct 2026	ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence.	ELRA catalogue, ICE-GB, British component of the International Corpus of English (ELRA-W0021) A catalogue or marketplace listing
La Depeche de Kabylie corpus (about 1,570,000 words of Amazigh newspaper text)	150.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, La Depeche de Kabylie corpus (ELRA-W0322) A catalogue or marketplace listing
Karl May Korpus (the works of the German author Karl May, about 1.6 million words)	3500.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence. One author's collected fiction, priced by the corpus; about $0.0025 a word.	ELRA catalogue, Karl May Korpus (ELRA-W0016) A catalogue or marketplace listing
MEMODATA tagged French text: a tagged corpus of 50 books (19th-century novels by Balzac, Hugo, Stendhal and others, plus legal texts)	3846.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence. Public-domain classics with morphological tagging, so the price is for the annotation, not the rights.	ELRA catalogue, MEMODATA tagged French text: a tagged corpus of 50 books (ELRA-W0012) A catalogue or marketplace listing
MULTEXT JOC corpus (written questions and answers of the EU Official Journal, about 5 million words in five languages)	5000.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, MULTEXT JOC corpus (ELRA-W0017) A catalogue or marketplace listing
PAROLE French corpus (20.1 million words: Le Monde, CNRS Editions books, periodicals)	4300.00 € per corpus, research use only	Oct 2026	ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence.	ELRA catalogue, PAROLE French corpus (ELRA-W0020) A catalogue or marketplace listing
PRESS 65 (one million running words of 1965 Swedish newspapers, annotated to sentence level)	20000.00 € per corpus, research use only	Oct 2026	ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence. An outlier: the highest per-word list price in the sweep, for a small, old, research-only corpus.	ELRA catalogue, PRESS 65 (ELRA-W0010) A catalogue or marketplace listing
Wojood nested Arabic named-entity corpus (about 550,000 tokens, MSA and dialect, annotated)	3750.00 € per corpus, commercial use	Oct 2026	ELRA non-member price for a commercial-use (VAR) licence.	ELRA catalogue, Wojood nested Arabic named-entity corpus (ELRA-W0325) A catalogue or marketplace listing
76,087 full-text books in 10 Indian languages (29.3 million pages, about 11.7 billion tokens)	$526,300 per corpus of 76,087 books	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Custom licence, commercial AI use. Seventy times cheaper per title than the rights-cleared e-book collections on the same marketplace. Listed 11 Sep 2026.	Opendatabay, Multilingual Books Corpus, 10 Indian Languages A catalogue or marketplace listing
Academic theses (multi-domain, 641,727 theses and 7.3B+ words in the full set), price for an initial batch of 20,000 theses	$972,200 per initial batch of 20,000 theses	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. The price covers the first 20,000 theses only; larger batches are quoted. Listed 24 Jul 2026.	Opendatabay, 7.3B+ Words Multi-Domain Academic Thesis Dataset (Infobay.AI) A catalogue or marketplace listing
6,400+ Arabic non-STEM textbooks, 336M+ words (PDF), commercial AI licence	$506,800 per dataset of 6,400+ textbooks	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.	Opendatabay, 336M+ Words covering 6K+ Arabic Non-STEM Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
1,000+ Bengali non-STEM textbooks, 74M+ words (PDF), commercial AI licence	$108,600 per dataset of 74M+ words	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.	Opendatabay, 74M+ Words covering 1K+ Bengali Non-STEM Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
6,000+ English non-STEM textbooks, 521M+ words (PDF), commercial AI licence	$493,400 per dataset of 6,000+ textbooks	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 25 Jul 2026.	Opendatabay, 521M+ Words covering 6K+ English Non-STEM Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
8,939 English STEM textbooks, 855M+ words (PDF), commercial AI licence	$696,800 per dataset of 8,939 textbooks	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. The reseller says it sublicenses under agreements with its data vendors. Listed 25 Jul 2026.	Opendatabay, 855M+ Words covering 8K+ English STEM Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
2,800+ Hindi non-STEM textbooks, 250M+ words (PDF), commercial AI licence	$222,100 per dataset of 2,800+ textbooks	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.	Opendatabay, 250M+ Words covering 2K+ Hindi Non-STEM Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
5,169 Indonesian STEM textbooks, 208.30M+ words (PDF), commercial AI licence	$402,900 per dataset of 5,169 textbooks	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 25 Jul 2026.	Opendatabay, 208.30M+ Words covering 5K+ Indonesian STEM Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
Legal contracts (1M+ in the full set), price for an initial batch of 100,000 contract documents	$487,600 per initial batch of 100,000 legal documents	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 25 Jul 2026.	Opendatabay, 1M+ Legal Contracts Dataset (Infobay.AI) A catalogue or marketplace listing
300+ legal textbooks, 32M+ words across 80K+ pages (PDF), commercial AI licence	$26,600 per dataset of 300+ textbooks	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 26 Sep 2026.	Opendatabay, 300+ Legal Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
140+ Sanskrit non-STEM textbooks, 8M+ words (PDF), commercial AI licence	$11,400 per dataset of 8M+ words	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.	Opendatabay, 8M+ Words covering 140+ Sanskrit Non-STEM Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
1,000+ Urdu non-STEM textbooks, 57M+ words (PDF), commercial AI licence	$81,700 per dataset of 57M+ words	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.	Opendatabay, 57M+ Words covering 1K+ Urdu Non-STEM Textbooks Dataset (Infobay.AI) A catalogue or marketplace listing
TIDES Extraction (ACE) 2003 Multilingual Training Data (about 231,000 words in Arabic, Chinese and English annotated for entities and relations)	$3,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 16 February 2004.	LDC catalog, TIDES Extraction (LDC2004T09) A catalogue or marketplace listing
Annotated English Gigaword (nearly ten million newswire documents, over four billion words, with parses, named entities and coreference)	$6,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 November 2012. The same fee as the unannotated Fifth Edition: LDC charges nothing extra for the machine annotation layer.	LDC catalog, Annotated English Gigaword (LDC2012T21) A catalogue or marketplace listing
AQUAINT-2 Information-Retrieval Text Research Collection (about 907K English newswire documents, Oct 2004 to Mar 2006)	$500.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 19 December 2008.	LDC catalog, AQUAINT-2 Information-Retrieval Text Research Collection (LDC2008T25) A catalogue or marketplace listing
AQUAINT Corpus of English News Text (roughly 375 million words: AP, NYT and Xinhua newswire)	$300.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 26 September 2002.	LDC catalog, AQUAINT Corpus of English News Text (LDC2002T31) A catalogue or marketplace listing
Arabic Gigaword first edition (about 1 million news documents, 400 million words of Arabic)	$3,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 22 July 2003.	LDC catalog, Arabic Gigaword first edition (LDC2003T12) A catalogue or marketplace listing
Chinese English News Magazine Parallel Text (20 million Chinese characters with 9 million words of English translation)	$2,500.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 June 2005. Per word counts the English side only.	LDC catalog, Chinese English News Magazine Parallel Text (LDC2005T10) A catalogue or marketplace listing
Chinese Gigaword first edition (about 1 billion words of Mandarin news text)	$3,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 22 May 2003.	LDC catalog, Chinese Gigaword first edition (LDC2003T09) A catalogue or marketplace listing
Chinese Gigaword Second Edition (about 1.3 billion words of Chinese newswire)	$3,500.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 17 August 2005.	LDC catalog, Chinese Gigaword Second Edition (LDC2005T14) A catalogue or marketplace listing
Chinese News Translation Text Part 1 (474,000 Chinese characters with commissioned English translations, about 285,000 words)	$2,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 March 2005. Human-translated text prices 28 to 54 times the automatically extracted parallel sets (LDC2007T08, LDC2007T09).	LDC catalog, Chinese News Translation Text Part 1 (LDC2005T06) A catalogue or marketplace listing
Chinese Treebank 9.0 (about two million words of annotated and parsed Chinese text)	$300.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 June 2016.	LDC catalog, Chinese Treebank 9.0 (LDC2016T13) A catalogue or marketplace listing
English Gigaword first edition (about 1.8 billion words of English newswire, 4.1 million documents)	$3,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 28 January 2003.	LDC catalog, English Gigaword first edition (LDC2003T05) A catalogue or marketplace listing
English Gigaword Second Edition (about 2.3 billion words of English newswire)	$3,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 July 2005.	LDC catalog, English Gigaword Second Edition (LDC2005T12) A catalogue or marketplace listing
English Gigaword Fifth Edition (4.03 billion words, 9.88 million newswire documents, seven English news services, 1994 to 2010)	$6,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 17 June 2011. Word count is LDC's K-wrds total (whitespace tokens, in thousands).	LDC catalog, English Gigaword Fifth Edition (LDC2011T07) A catalogue or marketplace listing
English Gigaword Fifth Edition, commercial reduced-licence fee (4.03 billion words of newswire)	$3,000.00 Reduced-License per corpus (commercial licence for for-profit LDC members, plus membership)	Oct 2026	Reduced-licence fee: the commercial-use price LDC charges for-profit members for corpora from closed membership years, on top of the membership fee. Corpus released 17 June 2011.	LDC catalog, English Gigaword Fifth Edition, commercial reduced-licence fee (LDC2011T07) A catalogue or marketplace listing
English Web Treebank (254,830 words of weblogs, newsgroups, email, reviews and Q&A, hand-annotated for syntax)	$175.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 16 August 2012.	LDC catalog, English Web Treebank (LDC2012T13) A catalogue or marketplace listing
LDC for-profit standard membership: access to 16 corpora published in the membership year (text and speech)	Standard membership: $34,000 per membership year (16 corpora)	Oct 2026	The for-profit membership is what grants commercial technology-development rights; the subscription tier ($40,000) receives every corpus published that year. Not-for-profits pay $2,400.	LDC, Join LDC (membership fees) A catalogue or marketplace listing
French Gigaword Third Edition (863 million words, 3.16 million newswire documents)	$4,500.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 September 2011.	LDC catalog, French Gigaword Third Edition (LDC2011T10) A catalogue or marketplace listing
ISI Arabic-English Automatically Extracted Parallel Text (1,124,609 sentence pairs, about 31M English words)	$4,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 20 February 2007. Per word counts the English side only.	LDC catalog, ISI Arabic-English Automatically Extracted Parallel Text (LDC2007T08) A catalogue or marketplace listing
ISI Chinese-English Automatically Extracted Parallel Text (558,567 sentence pairs, about 16M English words)	$4,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 18 April 2007. Per word counts the English side only.	LDC catalog, ISI Chinese-English Automatically Extracted Parallel Text (LDC2007T09) A catalogue or marketplace listing
Prague Czech-English Dependency Treebank 2.0 (over two million words in close to 100,000 sentences)	$100.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 June 2012.	LDC catalog, Prague Czech-English Dependency Treebank 2.0 (LDC2012T08) A catalogue or marketplace listing
Spanish Gigaword Second Edition (1.09 billion words, 3.7 million newswire documents)	$4,000.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 17 July 2009.	LDC catalog, Spanish Gigaword Second Edition (LDC2009T21) A catalogue or marketplace listing
Treebank-2 (over 1.6 million words of hand-parsed Dow Jones and other text)	$1,500.00 Non-Member per corpus (research-only licence)	Oct 2026	Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus first released 1995 (catalog LDC95T7).	LDC catalog, Treebank-2 (LDC95T7) A catalogue or marketplace listing
5 million real consumer reviews and complaints with intent labels (PissedConsumer data)	Average price per record/row is $0,15 per record (review)	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. The full 5M set lists at $736,800; final price depends on languages and licensing scope. Listed 7 Jul 2026.	Opendatabay, 5M Consumer reviews & Complaints for NLP Training Data A catalogue or marketplace listing
High-quality Arabic web corpus (FineWeb-edu Common Crawl text machine-translated by an LLM), 13.3M documents, 14.9B tokens	$15,800 per corpus of 14,895,034,936 tokens	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Derived from Common Crawl and translated by machine, so synthetic text under a custom licence; listed 17 Oct 2025.	Opendatabay, High Quality Arabic Corpus (Token Haven) A catalogue or marketplace listing
High-quality Spanish, Arabic and Norwegian web corpus (FineWeb-edu Common Crawl text machine-translated by an LLM), about 40M documents, 45B tokens	$36,800 per corpus of 45B tokens	Oct 2026	Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. The bundle of the three single-language corpora; listed 20 Oct 2025.	Opendatabay, High Quality Multilingual Corpus (Token Haven) A catalogue or marketplace listing
GPU cost to machine-translate 13.4M FineWeb-edu web documents into Norwegian with an LLM (15.3B tokens)	15,000 H100 Hours per language (~$36000) per language (15,263,569,765 tokens)	Oct 2026	The producer's own stated compute cost; it sells the same corpus for $15,800, so synthetic translated text lists below its own production cost.	Opendatabay, High Quality Norwegian Corpus (Token Haven) A vendor's published price guide
Times of India and Economic Times content (Bennett, Coleman & Co) surfaced in ChatGPT under an OpenAI partnership	around $5 million annually per year	Sep 2026	Industry-source estimate; neither party disclosed terms. Shows what a large non-US news group gets versus $20M to $50M for top US brands.	exchange4media Press report
Indian Express Group content licensed to OpenAI for ChatGPT	about $3 million a year per year	Sep 2026	Industry-source estimate; terms undisclosed by the parties.	exchange4media Press report
News analysts, reporters and journalists evaluating AI-generated news content (4+ years' experience)	$70-$110 per hour per hour	Aug 2026	What a working journalist is paid by the hour to grade AI news output, against nothing per article when a publisher licenses their archive.	Mercor job board, News analysts, reporters and journalists evaluating AI-generated news content The payer's own page
Reddit other revenue (mostly data licensing, plus products sold directly to users), second quarter of 2026, all licensees	$43 million, up 24% year over year per quarter (all licensees)	Jul 2026	Aggregate, not one deal: Reddit's 10-Q says substantially all its licensing contract value comes from two partners, unnamed. $43 million of $805 million total revenue is about 5% (our arithmetic; the 10-Q gives $43,280 thousand of $804,905 thousand).	Reddit Q2 2026 results, Exhibit 99.1 (SEC) The payer's own page
Total damages awarded on appeal to 78 rights holders of Communication Books titles copied into the National Institute of Korean Language written corpus (Woongjin and Woongjin Booxen), 2026	약 344만 원 (about 3.44 million won) total award (78 rights holders)	Jul 2026	About 62 percent of the roughly 5.5 million won claimed. Separate from the 2023 settlement between the Korean Publishers Association, the National Institute of Korean Language and Woongjin Booxen, whose amount was not disclosed.	Bizhankook, Woongjin loses AI copyright lawsuit (5 August 2026) Press report
Per-title damages the Seoul Central District Court (appeal) set for e-books copied without permission into the National Institute of Korean Language written corpus, 78 rights holders of Communication Books titles against Woongjin and Woongjin Booxen, 2026	전자책 정가의 70%에 인쇄 부수 5부 (70% of the e-book list price, times 5 copies, times the plaintiff's share) per title, as a multiple of the e-book list price	Jul 2026	A court-set price for unlicensed use of a book in a training corpus: 3.5 times the e-book list price per title, times each plaintiff's share; the court held the use was not fair use. The appeal upheld a first-instance ruling of February 2025. The judgment itself was not read.	Bizhankook, Woongjin loses AI copyright lawsuit (5 August 2026) Press report
Bartz v. Anthropic class settlement: the $1.5 billion fund divided over the 482,460 works on the final Works List (books downloaded from pirate libraries)	$1.5 billion for 482,460 Works per settlement covering 482,460 works	Jul 2026	Gross settlement value per listed work, from the final approval order; the court itself rounds it to 'approximately $3,000'. It pays for past acquisition of pirated copies (the court found training itself fair use), so it prices liability, not a training licence going forward.	Bartz v. Anthropic, Order Granting Final Approval of Class Action Settlement (N.D. Cal., Dkt. 680, 20 July 2026) Court record
Bartz v. Anthropic class settlement: the fund left after the court-awarded fees, expenses, cost reserve and service awards, divided over the 440,490 works claimed by April 2026	$1.5 billion less $101,561,111 fees, $2,635,197.46 expenses, $18,220,000 cost reserve and $45,000 service awards, over 440,490 claimed works net settlement fund for 440,490 claimed works	Jul 2026	Because the fund is non-reversionary and paid pro rata per claimed work, the 8.7 percent of works never claimed raise the share of those claimed: net of the court's cut of class counsel to about 6.8 percent, a claimed work carries about $3,127 before the author and publisher split it (by default 50-50 for trade books). First payments were expected on or before 15 November 2026.	Bartz v. Anthropic, Order Granting Final Approval of Class Action Settlement (N.D. Cal., Dkt. 680, 20 July 2026) Court record
Books in the Bartz v. Anthropic class settlement (pirated copies used for training)	approximately $3,000 per work, from a $1.5 billion fund per book	Jul 2026	A litigation settlement for past unlicensed use, not a licence price; before fees, split between author and publisher.	Authors Guild (on the court's final approval) Court record
Wiley scholarly and learning content licensed for AI (all licensees, fiscal year to 30 April 2026)	$49.1 million (prior year $40 million) per year (all AI licensees)	Apr 2026	Aggregate revenue, not one deal; $19.4M of it came from other publishers' content Wiley resold for a royalty.	John Wiley & Sons Form 10-K (SEC) The payer's own page
News Corp US and UK news content, including archives, licensed to Meta for AI	up to US$50 million a year, at least three years per year	Mar 2026	Terms from people familiar; News Corp confirmed the deal but not the money.	Mediaweek (citing The Wall Street Journal) Press report
What one academic author received as his share when a small, ethical academic publisher licensed one of his monographs to an unnamed AI training company (opt-in by tick box, publisher keeps 50% of the total fee, buyer undisclosed under NDA)	£500 per title (author's cheque after the publisher's 50% cut, one-off)	Feb 2026	First-hand account by the payee of a payment received at the end of 2025. The publisher and the AI company are not named (the publisher was under an NDA), so this is one self-reported data point, but it is an opt-in programme at a 50/50 split, like HarperCollins. The author calls the amount 'rather pathetically small' and the most he has made from writing in a year.	Richard Pinner, Uniliterate, Academic Publishing AI Deals: My Biggest Paycheque Ever (24 February 2026) The payer's own page
Author's default half of the per-work award in the Bartz v. Anthropic settlement (trade and university press books on the Works List)	approximately $3,000 per work, by default split 50-50 between authors and publishers per work (author's default 50% share)	Jan 2026	Damages for past use of pirated copies, not a licence price; gross of fees (the court later cut class counsel's award to about 6.8% of the fund). Education works have no default split. The gross $3,000 is on the board as bartz-anthropic-per-work.	Bartz v. Anthropic, court-approved long-form class notice (updated deadline 8 Jan 2026) Court record
High end of the per-book AI licensing offers Created by Humans says it has seen for authors' books (founder's statement, no buyer named)	tens of thousands per book per book	Nov 2025	An intermediary's own claim about offers it has received, with no buyer, title or deal named, so it is an upper-end signal only. Created by Humans says it takes a 20% fee and declines 'low ball offers' without naming their level.	Written Word Media podcast, AI Rights and the Future of Publishing with Created by Humans founder Trip Adler Press report
AI training royalty per title paid to History Press (Arcadia Publishing) authors who stayed in Arcadia's 2025 sublicence to an unnamed technology company; one author took it for three titles	$1,020, or $340 per book per title (author's royalty, one-time use)	Nov 2025	Author-side money in a sublicence that went ahead (inclusion was the default and this author accepted for all three titles). Two other authors were offered the same $340; Arcadia's own fee and the buyer are undisclosed. Arcadia has a catalogue of nearly 17,000 titles.	Slate, Arcadia Publishing: Your town's local history books might be used to train A.I. Press report
One-time AI training royalty Arcadia Publishing offered a Belt Publishing author (Rethinking Fandom, 2022) in its second round of sublicensing to an unnamed major technology company, opt-out by default	$355 per title (author's royalty, one-time use)	Oct 2025	Arcadia's email reprinted in full by the author, who has one book with Belt and opted out. Opt-out deadline 7 November 2025, payment expected in spring 2026. Belt's director said the second-round titles were finalized in late October 2025.	Cup of Coffee newsletter (Craig Calcaterra), 3 November 2025, reprint of Arcadia Publishing's 'Royalty Opportunity' email The payer's own page
Bartz v. Anthropic class settlement: the payment Anthropic owes for each class work added to the Works List above 500,000	$3,000/book for each Class work above 500,000 per additional book	Sep 2025	The settlement's marginal price per book, written into the agreement: the fund of $1.5 billion rises by $3,000 for each work above 500,000. The final Works List stopped at 482,460, so the clause was not triggered; it is the per-work rate the parties agreed, not a payment made.	Bartz v. Anthropic, Class Action Settlement Agreement (N.D. Cal. 3:24-cv-05417, Dkt. 363-3), on the settlement website Court record
Publisher content used in Perplexity's Comet browser and assistant (revenue share)	80% of $5 monthly Comet Plus revenue; $42.5 million pool for early partners total pool (plus 80% of subscription revenue)	Aug 2025	A usage-based revenue share, not a flat licence; Perplexity keeps 20% to cover compute.	Axios Press report
Royalty Bloomsbury offers academic and professional authors who opt in to AI training licences of their backlist titles	a royalty rate of 20 percent share of licence revenue per title	Aug 2025	A split, not a dollar figure: the publisher keeps 80%, against 50% at HarperCollins. Bloomsbury's 2025/26 interim results report its first non-exclusive AI licensing agreement without a fee.	Publishers Lunch, Bloomsbury explores AI licensing Press report
New York Times news, cooking and Athletic content licensed to Amazon for AI training and summaries	between $20 million to $25 million per year per year	Jul 2025	First AI licence for the Times; figure from people familiar, not confirmed by either company.	TheWrap (citing The Wall Street Journal) Press report
Author's expected return when Johns Hopkins University Press licenses a backlist title to an unnamed LLM developer (opt-out programme, about 3,000 backlist titles)	less than $100 per title per license per title, per licence (author's share)	Jul 2025	The press director's own guidance to authors; the buyer and the press's total fee were not disclosed. An upper bound only, and about 1/25th of the HarperCollins author share.	The Baltimore Banner, Johns Hopkins Press will license its books to train AI Press report
Taylor & Francis (Informa) incremental non-recurring AI data licensing revenue in the first half of 2025, including a new 2025 contract with a different, undisclosed LLM customer (academic books and content)	about a £15 million incremental, non-recurring data licensing revenue number in the first half revenue in the half year to 30 June 2025	Jul 2025	The seller's own finance director on its results call. Informa would not name the customer or the terms ('fairly under wraps'); its 2025 annual report adds only that it signed a further data access contract with another AI technology provider and that non-recurring data revenue was lower than 2024's $75m+.	Informa plc, Half Year Results 2025, transcript of the results call (23 July 2025) The payer's own page
One-time AI training royalty Arcadia Publishing offered the author of a local-history title (Images of Baseball: Kane County Cougars) for a sublicence to an unnamed major technology company, first round, opt-out by default	$205 per title (author's royalty, one-time use)	Jul 2025	Arcadia's own email, posted as a screenshot by the author; the buyer and Arcadia's own fee were not disclosed. Inclusion was the default ('No action is needed to be included'), opt-out by 18 July 2025. Slate (3 November 2025) reports the same round paid other History Press authors $340 a title.	Bluesky, David Malamut's post of Arcadia Publishing's email to authors (screenshot) The payer's own page
What one Taylor & Francis author says they earned for their book under the publisher's AI training licence to Microsoft	only $97 per book (author's royalty)	Jun 2025	One anonymous author's receipt, quoted by Bloomberg (paywalled) and repeated by Slate; Taylor & Francis pays authors under their existing contract royalty terms and gave no opt-out. Close to the under-$100 per title Johns Hopkins University Press told its authors to expect.	Slate (3 November 2025), citing Bloomberg, Academic publishers sign AI deals as Trump cuts research funding (10 June 2025) Press report
HarperCollins AI training licence fee actually paid to one author for one nonfiction backlist title (Writing Without Bullshit, 2016): three-year opt-in licence, author signed and received a cheque about a month later	$2,500 per title (author's share of the per-title fee, before the agent's 15%)	Apr 2025	First-hand account by the payee that the money was paid: he signed and a month or so later received the cheque (shown as an image on the page). It confirms one author was paid, not how many authors opted in or were paid. The buyer is not named by HarperCollins; Bernoff calls it 'this AI company (likely Microsoft)', so Microsoft is his inference. The programme is the one already held as an offer (harpercollins-microsoft-author-share-per-title) and as the gross (book-title); this is the realised payment.	Josh Bernoff, Why I took the $2,500 check to license my book for AI (bernoff.com; withoutbullshit.com redirects to it) The payer's own page
Black Inc (Melbourne trade publisher): contract variation asking its authors to grant AI training rights to their books, author's share of the licence receipts	net receipts split with the author 50/50 share of net licence receipts per title	Mar 2025	The publisher's offer to its own authors, not a buyer's price: no licensee or fee per title was disclosed. The Australian Society of Authors called 50/50 unfair and backs 75 percent to the author, as the Authors Guild does.	The Guardian (Australia): Black Inc asks writers to sign AI agreements Press report
Author's half of HarperCollins' per-title AI training fee (select nonfiction backlist, three-year licence, buyer reported as Microsoft)	$2,500 per book per book (author's share of $5,000)	Feb 2025	What the author receives: the publisher keeps the other $2,500. Opt-in; outputs capped at 200 consecutive words or 5% of a book. The $5,000 gross is already on the board as book-title.	The Decoder, Microsoft offers authors $5,000 to train AI on their books Press report
Royalty Cambridge University Press pays authors who opt in when their work is used in a generative AI licensing deal	a royalty of 20% share of licence revenue per title	Feb 2025	A split, not a dollar figure; per the post, licence revenue is divided equally across all content items in the licensed collection before the 20% applies.	Silicon Reckoner (Michael Harris), quoting Cambridge University Press's letter to authors Press report
Share of AI licensing income that some academic presses offer their authors per title, as reported by the Authors Guild (presses not named)	a mere 25 percent share of AI licence income per title	Dec 2024	A split, not a dollar figure, and the presses are unnamed. It sits between Bloomsbury and Cambridge (20%) and HarperCollins and Black Inc (50%); the Guild calls 75 to 85 percent to the author fair.	Authors Guild, AI Licensing for Authors: Who Owns the Rights and What's a Fair Split? Press report
One nonfiction book title, licensed for training for three years (HarperCollins)	$5,000 per title per title	Nov 2024	A signed deal between HarperCollins and an AI company reported as Microsoft, paid for each backlist title whose author opts in, split equally between author and publisher. Agents called it a one-off done by contract addenda.	Publishers Weekly Press report
Wiley's learning and research content licensed to tech companies for AI, total earned to date by August 2024	$23 million earned to date from AI deals	Aug 2024	Wiley's first AI licensing money, with a further $21 million then expected in the same financial year; authors were given no opt-out. Wiley's later fiscal years are on the board as wiley-fy2026-ai-licensing.	infoDOCKET, citing The Bookseller Press report
Dotdash Meredith (now People Inc.) magazine and web content licensed to OpenAI (2024 deal)	reported around 16 million dollars a year per year	May 2024	A reported figure; the publisher has described the deal as a guaranteed fixed payment plus a variable part tied to display in ChatGPT.	LLM Pulse, AI content licensing deals 2023 to 2026 Press report
News Corp's current and archived journalism licensed to OpenAI for five years (2024 deal)	reported 250 million dollars over five years over five years	May 2024	The largest known publisher deal, reported and not confirmed by the parties; it covers training and display in ChatGPT. The per-year figure is the total divided evenly by the term.	LLM Pulse, AI content licensing deals 2023 to 2026 Press report
Reddit's user posts licensed to OpenAI for training and display (2024 deal)	reported around 70 million a year per year	May 2024	A reported figure the parties have not confirmed; the deal also covers real-time display in ChatGPT, so not all of it pays for training.	LLM Pulse, AI content licensing deals 2023 to 2026 Press report
Taylor & Francis academic books and journals licensed to Microsoft for AI (first year of the deal)	almost £8m ($10m) in its first year first year of the deal	May 2024	The $10m is the source's own conversion of about £8m. Authors were not told, could not opt out and were paid nothing extra, which is the rights risk a buyer prices in.	ISSN International Centre newsletter, Taylor & Francis sells access to research to Microsoft AI Press report
General text sold through a data marketplace (Defined.ai's quoted rate)	$0.001 a word per word	Apr 2024	$1,000 per million words. Edited, published work selected by a buyer has licensed for fifty times that.	Reuters, via The Decoder Press report
Financial Times journalism licensed to OpenAI (2024 deal)	reported 5 to 10 million dollars a year per year	Apr 2024	A reported range the parties have not confirmed. Covers training plus display with attribution in ChatGPT.	LLM Pulse, AI content licensing deals 2023 to 2026 Press report
Reddit's user posts licensed to Google for AI training and access (2024 deal)	a reported 60 million a year per year	Feb 2024	A reported figure the parties have not confirmed. Reddit's own IPO filing disclosed $203.0 million of aggregate contract value across its January 2024 data licences, over two to three years.	LLM Pulse, AI content licensing deals 2023 to 2026 Press report
Reddit's data licences signed in January 2024 (access to search and analyse Reddit data), aggregate of several contracts	$203.0 million aggregate contract value, terms of two to three years aggregate contract value of several licences, two to three years each	Jan 2024	The seller's own filing, the firmest figure behind the reported Google and OpenAI deals. It sums several contracts; Reddit expected at least $66.4 million of it as revenue in 2024.	Reddit, Form S-1 registration statement (SEC) The payer's own page
Apple's reported multiyear offers to news publishers (Conde Nast, NBC News, IAC) for archives to train its generative AI	Reported offers of 50M-plus per publisher per publisher (multiyear, total)	Dec 2023	Offers in talks first reported by The New York Times in December 2023; no signed deal has been confirmed, so this is the buyer's opening offer, not a licence.	LLM Pulse, AI content licensing deals map (updated 28 Sep 2026) Press report
Axel Springer news content licensed to OpenAI (December 2023 deal)	a reported figure in the tens of millions of euros per deal (term not stated)	Dec 2023	OpenAI's first global news publisher deal. Words only: the source gives an order of size, not a number, and does not say whether it is annual.	LLM Pulse, AI content licensing deals 2023 to 2026 Press report
Per-title settlement Woongjin Booxen set with the publishers whose Bookopia e-books it used without permission in the National Institute of Korean Language written corpus (an AI and language-research corpus), Korea, 2022	전자책 정가의 70%의 3copy (70% of the e-book list price, times 3 copies) per title, as a multiple of the e-book list price	Sep 2022	A formula, not an amount: about 2.1 times the e-book list price per title. Woongjin called it a settlement of fees, not compensation. Seoul Shinmun (26 August 2022) reports 6,194 of the 15,993 disputed titles were settled at the level paid to publishers under formal contracts.	Kyunghyang Shinmun, opinion column by culture researcher Lee Yung-hee, 13 September 2022 Press report
Korean government (National Institute of Korean Language) written-corpus collection project: total project cost over the 20,053 published works the contractor used from 1,226 publishers, 2019	총 사업비는 30억 9000만원 (3.09 billion won) for 2만 53종 (20,053 works) per project of 20,053 works	May 2019	A buyer paying a vendor to assemble a corpus, so the cost covers processing as well as rights; only 4,060 works were licensed from their publishers, and 15,933 Bookopia e-books were used without consent (later settled, see the related Korean rows). Bizhankook reports the contract required at least 80% of the fee to go to copyright clearance.	Seoul Shinmun, Woongjin Booxen infringed copyright in 16,000 titles (26 August 2022) Press report

Silicon-DPO Platinum, The Reasoning-Code Dataset: 5,001 preference pairs with think-trace and self-corrected code, verified by Python unit tests in a sandbox (coding 35%, finance 20%, STEM 15%; JSONL), Commercial / Startup licence, sold on Gumroad
Price
€499
per dataset of 5,001 pairs (one-time)
When
Oct 2026
What to know
The commercial tier of a two-tier listing by one individual seller; undated page, read 2026-10-17. Only 35% of the pairs are coding, so it is filed as a text preference dataset, not a code dataset, although the listing calls it a reasoning-code dataset. Gumroad's own page shows a USD price of 55.08 for the EUR 49 tier (1.124 USD per EUR); the ECB rate is used here. About ten times the research tier.
Source
Gumroad, Silicon-DPO Platinum: The Reasoning-Code Dataset (storefront ryanride)
A catalogue or marketplace listing
Silicon-DPO Platinum, The Reasoning-Code Dataset: 5,001 preference pairs with think-trace and self-corrected code, verified by Python unit tests in a sandbox (coding 35%, finance 20%, STEM 15%; JSONL), Personal / Research licence, sold on Gumroad
Price
€49
per dataset of 5,001 pairs (one-time)
When
Oct 2026
What to know
The non-commercial tier of the same listing, kept apart from commercial prices. Undated page, read 2026-10-17.
Source
Gumroad, Silicon-DPO Platinum: The Reasoning-Code Dataset (storefront ryanride)
A catalogue or marketplace listing
Humanities expert work on Mercor (standing talent network; band its listings have posted at)
Price
$45 to $60 an hour
per hour
When
Oct 2026
What to know
The posting closed; its page now shows the standing Mercor talent network for the field, which states the band its listings have paid and names a project's own rate only when it invites someone (was $50 an hour on the posting).
Source
Mercor, English Language and Literature Expert (closed) and its Humanities Expert talent network
The payer's own page
NLP entity annotation (named entity labelling), Label Your Data starting price
Price
$0.02 per entity for NLP tasks
per entity, starting price
When
Oct 2026
What to know
A floor ('starts as low as'), not a typical price; undated page read on 2026-10-14.
Source
Label Your Data, Pricing (FAQ)
A vendor's published price guide
AI language expert, French (native speakers with teaching experience) on Alignerr
Price
$25-30/hr
per hour
When
Oct 2026
What to know
The language tier sits well below expert and coding tiers on the same board ($40 to $120 an hour).
Source
Alignerr, open roles
The payer's own page
100,000+ rights-cleared academic and professional e-books (PDF, EPUB; English, Hindi and other languages) from Arts & Science Academic Publications
Price
$53,514,000
per collection of 100,000+ e-books (non-exclusive, perpetual)
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. A publisher selling its own catalogue; listed 24 Jul 2026.
Source
Opendatabay, 100K+ EBooks related to STEM & Non-STEM in multiple languages
A catalogue or marketplace listing
18,000+ rights-cleared STEM and non-STEM EPUB e-books from Indian publishers, licensed by Arts & Science Academic Publications
Price
$9,473,700
per collection of 18,000+ e-books (non-exclusive, perpetual)
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 24 Jul 2026.
Source
Opendatabay, 18000+ EBooks in epub format STEM & Non-STEM
A catalogue or marketplace listing
Quora questions and answers dataset (10.2M+ records, 21 fields) scraped and sold by Bright Data
Price
Up to $0.0025 per record, Min Order $250
per record
When
Oct 2026
What to know
Scraped forum text; the buyer carries the rights question, which is why it prices like a collection fee rather than a licence.
Source
Bright Data, Quora datasets
A catalogue or marketplace listing
Wikipedia articles dataset (9.2M+ records, 11 fields) scraped and sold by Bright Data
Price
$250 for 100K records (approximately $0.0025 per record)
per 100,000 records (one-time)
When
Oct 2026
What to know
Entry price for a one-time snapshot of openly licensed text; the price is for collection and delivery, not rights.
Source
Bright Data, Wikipedia datasets
A catalogue or marketplace listing
60M+ words of editorially curated magazine text (about 90 issues, English, Hindi and Indian regional languages), 100% cleared IP
Price
$150,000
per corpus of 60M+ words
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. A single publisher's own archive (Prakhar Goonj Publications), proprietary licence. Listed 2 Oct 2026.
Source
Opendatabay, 60M+ Words Multilingual AI/LLM Text Corpora (100% Cleared IP)
A catalogue or marketplace listing
Al-Hayat Arabic newspaper corpus (42,591 articles in 7 domains, 18.6M tokens)
Price
1440.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, Al-Hayat Arabic newspaper corpus (ELRA-W0030)
A catalogue or marketplace listing
Amharic-English bilingual news and legal corpus (232,653 Amharic and 291,701 English words)
Price
4000.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence. Per word counts both sides (232,653 + 291,701). A low-resource language pair prices about 100 times the per-word rate of big newswire sets.
Source
ELRA catalogue, Amharic-English bilingual news and legal corpus (ELRA-W0074)
A catalogue or marketplace listing
An-Nahar Arabic newspaper text, 1995 to 2000 (6 years, 45,000 articles and 24 million words a year)
Price
4788.00 €
per corpus, research use only
When
Oct 2026
What to know
ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence. Word count is 6 years x 24 million words as stated; one year alone is 1,008 EUR for a commercial organisation.
Source
ELRA catalogue, An-Nahar Arabic newspaper text, 1995 to 2000 (ELRA-W0027)
A catalogue or marketplace listing
Annotated tweet corpus in Arabizi, French and English (585,163 tweets, opinion and theme annotations)
Price
10000.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, Annotated tweet corpus in Arabizi, French and English (ELRA-W0323)
A catalogue or marketplace listing
Chinese-Vietnamese Parallel Corpus (200,000 sentence pairs, average 15 words per sentence)
Price
2100.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, Chinese-Vietnamese Parallel Corpus (ELRA-W0312)
A catalogue or marketplace listing
CRATER corpus (aligned telecom manuals, 1,000,000 tokens per language in English, French and Spanish)
Price
100.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence. Per word counts 3 languages x 1,000,000 tokens.
Source
ELRA catalogue, CRATER corpus (ELRA-W0003)
A catalogue or marketplace listing
Dutch PAROLE Distributable Corpus (3 million words of books, newspapers and periodicals)
Price
2500.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, Dutch PAROLE Distributable Corpus (ELRA-W0019)
A catalogue or marketplace listing
ECI/MCI European Corpus Initiative Multilingual Corpus I (over 98 million words in most major European languages)
Price
50.00 €
per corpus, research use only
When
Oct 2026
What to know
ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence. A distribution fee rather than a value price.
Source
ELRA catalogue, ECI/MCI European Corpus Initiative Multilingual Corpus I (ELRA-W0004)
A catalogue or marketplace listing
Archives of El Mundo newspaper, 2020 to 2022 (45,658 Spanish articles, El Mundo and sister titles)
Price
8850.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence. The rare case of a recent newspaper archive with a public commercial price; members pay 5,900 EUR.
Source
ELRA catalogue, Archives of El Mundo newspaper, 2020 to 2022 (ELRA-W0332)
A catalogue or marketplace listing
Archives of El Mundo newspaper, year 2020 (15,073 Spanish articles)
Price
3000.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, Archives of El Mundo newspaper, year 2020 (ELRA-W0333)
A catalogue or marketplace listing
Archives of El Mundo newspaper, year 2021 (14,461 Spanish articles)
Price
3000.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, Archives of El Mundo newspaper, year 2021 (ELRA-W0334)
A catalogue or marketplace listing
Archives of El Mundo newspaper, year 2022 (16,124 Spanish articles)
Price
3000.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, Archives of El Mundo newspaper, year 2022 (ELRA-W0335)
A catalogue or marketplace listing
English-Vietnamese Parallel Corpus (1,000,000 sentence pairs, average 20 words per sentence)
Price
12000.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence. About $0.0003 per word if both sides average 20 words.
Source
ELRA catalogue, English-Vietnamese Parallel Corpus (ELRA-W0311)
A catalogue or marketplace listing
Monolingual Greek corpus (1 million words of 1996 Eleftherotipia newspaper articles)
Price
600.00 €
per corpus, research use only
When
Oct 2026
What to know
ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence.
Source
ELRA catalogue, Monolingual Greek corpus (ELRA-W0014)
A catalogue or marketplace listing
ICE-GB, British component of the International Corpus of English (a million words of spoken and written English, fully parsed)
Price
1500.00 €
per corpus, research use only
When
Oct 2026
What to know
ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence.
Source
ELRA catalogue, ICE-GB, British component of the International Corpus of English (ELRA-W0021)
A catalogue or marketplace listing
La Depeche de Kabylie corpus (about 1,570,000 words of Amazigh newspaper text)
Price
150.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, La Depeche de Kabylie corpus (ELRA-W0322)
A catalogue or marketplace listing
Karl May Korpus (the works of the German author Karl May, about 1.6 million words)
Price
3500.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence. One author's collected fiction, priced by the corpus; about $0.0025 a word.
Source
ELRA catalogue, Karl May Korpus (ELRA-W0016)
A catalogue or marketplace listing
MEMODATA tagged French text: a tagged corpus of 50 books (19th-century novels by Balzac, Hugo, Stendhal and others, plus legal texts)
Price
3846.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence. Public-domain classics with morphological tagging, so the price is for the annotation, not the rights.
Source
ELRA catalogue, MEMODATA tagged French text: a tagged corpus of 50 books (ELRA-W0012)
A catalogue or marketplace listing
MULTEXT JOC corpus (written questions and answers of the EU Official Journal, about 5 million words in five languages)
Price
5000.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, MULTEXT JOC corpus (ELRA-W0017)
A catalogue or marketplace listing
PAROLE French corpus (20.1 million words: Le Monde, CNRS Editions books, periodicals)
Price
4300.00 €
per corpus, research use only
When
Oct 2026
What to know
ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence.
Source
ELRA catalogue, PAROLE French corpus (ELRA-W0020)
A catalogue or marketplace listing
PRESS 65 (one million running words of 1965 Swedish newspapers, annotated to sentence level)
Price
20000.00 €
per corpus, research use only
When
Oct 2026
What to know
ELRA non-member price for a commercial organisation under a non-commercial (research-only) licence. An outlier: the highest per-word list price in the sweep, for a small, old, research-only corpus.
Source
ELRA catalogue, PRESS 65 (ELRA-W0010)
A catalogue or marketplace listing
Wojood nested Arabic named-entity corpus (about 550,000 tokens, MSA and dialect, annotated)
Price
3750.00 €
per corpus, commercial use
When
Oct 2026
What to know
ELRA non-member price for a commercial-use (VAR) licence.
Source
ELRA catalogue, Wojood nested Arabic named-entity corpus (ELRA-W0325)
A catalogue or marketplace listing
76,087 full-text books in 10 Indian languages (29.3 million pages, about 11.7 billion tokens)
Price
$526,300
per corpus of 76,087 books
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Custom licence, commercial AI use. Seventy times cheaper per title than the rights-cleared e-book collections on the same marketplace. Listed 11 Sep 2026.
Source
Opendatabay, Multilingual Books Corpus, 10 Indian Languages
A catalogue or marketplace listing
Academic theses (multi-domain, 641,727 theses and 7.3B+ words in the full set), price for an initial batch of 20,000 theses
Price
$972,200
per initial batch of 20,000 theses
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. The price covers the first 20,000 theses only; larger batches are quoted. Listed 24 Jul 2026.
Source
Opendatabay, 7.3B+ Words Multi-Domain Academic Thesis Dataset (Infobay.AI)
A catalogue or marketplace listing
6,400+ Arabic non-STEM textbooks, 336M+ words (PDF), commercial AI licence
Price
$506,800
per dataset of 6,400+ textbooks
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.
Source
Opendatabay, 336M+ Words covering 6K+ Arabic Non-STEM Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
1,000+ Bengali non-STEM textbooks, 74M+ words (PDF), commercial AI licence
Price
$108,600
per dataset of 74M+ words
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.
Source
Opendatabay, 74M+ Words covering 1K+ Bengali Non-STEM Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
6,000+ English non-STEM textbooks, 521M+ words (PDF), commercial AI licence
Price
$493,400
per dataset of 6,000+ textbooks
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 25 Jul 2026.
Source
Opendatabay, 521M+ Words covering 6K+ English Non-STEM Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
8,939 English STEM textbooks, 855M+ words (PDF), commercial AI licence
Price
$696,800
per dataset of 8,939 textbooks
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. The reseller says it sublicenses under agreements with its data vendors. Listed 25 Jul 2026.
Source
Opendatabay, 855M+ Words covering 8K+ English STEM Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
2,800+ Hindi non-STEM textbooks, 250M+ words (PDF), commercial AI licence
Price
$222,100
per dataset of 2,800+ textbooks
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.
Source
Opendatabay, 250M+ Words covering 2K+ Hindi Non-STEM Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
5,169 Indonesian STEM textbooks, 208.30M+ words (PDF), commercial AI licence
Price
$402,900
per dataset of 5,169 textbooks
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 25 Jul 2026.
Source
Opendatabay, 208.30M+ Words covering 5K+ Indonesian STEM Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
Legal contracts (1M+ in the full set), price for an initial batch of 100,000 contract documents
Price
$487,600
per initial batch of 100,000 legal documents
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 25 Jul 2026.
Source
Opendatabay, 1M+ Legal Contracts Dataset (Infobay.AI)
A catalogue or marketplace listing
300+ legal textbooks, 32M+ words across 80K+ pages (PDF), commercial AI licence
Price
$26,600
per dataset of 300+ textbooks
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 26 Sep 2026.
Source
Opendatabay, 300+ Legal Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
140+ Sanskrit non-STEM textbooks, 8M+ words (PDF), commercial AI licence
Price
$11,400
per dataset of 8M+ words
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.
Source
Opendatabay, 8M+ Words covering 140+ Sanskrit Non-STEM Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
1,000+ Urdu non-STEM textbooks, 57M+ words (PDF), commercial AI licence
Price
$81,700
per dataset of 57M+ words
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Listed 27 Jul 2026.
Source
Opendatabay, 57M+ Words covering 1K+ Urdu Non-STEM Textbooks Dataset (Infobay.AI)
A catalogue or marketplace listing
TIDES Extraction (ACE) 2003 Multilingual Training Data (about 231,000 words in Arabic, Chinese and English annotated for entities and relations)
Price
$3,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 16 February 2004.
Source
LDC catalog, TIDES Extraction (LDC2004T09)
A catalogue or marketplace listing
Annotated English Gigaword (nearly ten million newswire documents, over four billion words, with parses, named entities and coreference)
Price
$6,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 November 2012. The same fee as the unannotated Fifth Edition: LDC charges nothing extra for the machine annotation layer.
Source
LDC catalog, Annotated English Gigaword (LDC2012T21)
A catalogue or marketplace listing
AQUAINT-2 Information-Retrieval Text Research Collection (about 907K English newswire documents, Oct 2004 to Mar 2006)
Price
$500.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 19 December 2008.
Source
LDC catalog, AQUAINT-2 Information-Retrieval Text Research Collection (LDC2008T25)
A catalogue or marketplace listing
AQUAINT Corpus of English News Text (roughly 375 million words: AP, NYT and Xinhua newswire)
Price
$300.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 26 September 2002.
Source
LDC catalog, AQUAINT Corpus of English News Text (LDC2002T31)
A catalogue or marketplace listing
Arabic Gigaword first edition (about 1 million news documents, 400 million words of Arabic)
Price
$3,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 22 July 2003.
Source
LDC catalog, Arabic Gigaword first edition (LDC2003T12)
A catalogue or marketplace listing
Chinese English News Magazine Parallel Text (20 million Chinese characters with 9 million words of English translation)
Price
$2,500.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 June 2005. Per word counts the English side only.
Source
LDC catalog, Chinese English News Magazine Parallel Text (LDC2005T10)
A catalogue or marketplace listing
Chinese Gigaword first edition (about 1 billion words of Mandarin news text)
Price
$3,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 22 May 2003.
Source
LDC catalog, Chinese Gigaword first edition (LDC2003T09)
A catalogue or marketplace listing
Chinese Gigaword Second Edition (about 1.3 billion words of Chinese newswire)
Price
$3,500.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 17 August 2005.
Source
LDC catalog, Chinese Gigaword Second Edition (LDC2005T14)
A catalogue or marketplace listing
Chinese News Translation Text Part 1 (474,000 Chinese characters with commissioned English translations, about 285,000 words)
Price
$2,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 March 2005. Human-translated text prices 28 to 54 times the automatically extracted parallel sets (LDC2007T08, LDC2007T09).
Source
LDC catalog, Chinese News Translation Text Part 1 (LDC2005T06)
A catalogue or marketplace listing
Chinese Treebank 9.0 (about two million words of annotated and parsed Chinese text)
Price
$300.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 June 2016.
Source
LDC catalog, Chinese Treebank 9.0 (LDC2016T13)
A catalogue or marketplace listing
English Gigaword first edition (about 1.8 billion words of English newswire, 4.1 million documents)
Price
$3,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 28 January 2003.
Source
LDC catalog, English Gigaword first edition (LDC2003T05)
A catalogue or marketplace listing
English Gigaword Second Edition (about 2.3 billion words of English newswire)
Price
$3,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 July 2005.
Source
LDC catalog, English Gigaword Second Edition (LDC2005T12)
A catalogue or marketplace listing
English Gigaword Fifth Edition (4.03 billion words, 9.88 million newswire documents, seven English news services, 1994 to 2010)
Price
$6,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 17 June 2011. Word count is LDC's K-wrds total (whitespace tokens, in thousands).
Source
LDC catalog, English Gigaword Fifth Edition (LDC2011T07)
A catalogue or marketplace listing
English Gigaword Fifth Edition, commercial reduced-licence fee (4.03 billion words of newswire)
Price
$3,000.00 Reduced-License
per corpus (commercial licence for for-profit LDC members, plus membership)
When
Oct 2026
What to know
Reduced-licence fee: the commercial-use price LDC charges for-profit members for corpora from closed membership years, on top of the membership fee. Corpus released 17 June 2011.
Source
LDC catalog, English Gigaword Fifth Edition, commercial reduced-licence fee (LDC2011T07)
A catalogue or marketplace listing
English Web Treebank (254,830 words of weblogs, newsgroups, email, reviews and Q&A, hand-annotated for syntax)
Price
$175.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 16 August 2012.
Source
LDC catalog, English Web Treebank (LDC2012T13)
A catalogue or marketplace listing
LDC for-profit standard membership: access to 16 corpora published in the membership year (text and speech)
Price
Standard membership: $34,000
per membership year (16 corpora)
When
Oct 2026
What to know
The for-profit membership is what grants commercial technology-development rights; the subscription tier ($40,000) receives every corpus published that year. Not-for-profits pay $2,400.
Source
LDC, Join LDC (membership fees)
A catalogue or marketplace listing
French Gigaword Third Edition (863 million words, 3.16 million newswire documents)
Price
$4,500.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 September 2011.
Source
LDC catalog, French Gigaword Third Edition (LDC2011T10)
A catalogue or marketplace listing
ISI Arabic-English Automatically Extracted Parallel Text (1,124,609 sentence pairs, about 31M English words)
Price
$4,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 20 February 2007. Per word counts the English side only.
Source
LDC catalog, ISI Arabic-English Automatically Extracted Parallel Text (LDC2007T08)
A catalogue or marketplace listing
ISI Chinese-English Automatically Extracted Parallel Text (558,567 sentence pairs, about 16M English words)
Price
$4,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 18 April 2007. Per word counts the English side only.
Source
LDC catalog, ISI Chinese-English Automatically Extracted Parallel Text (LDC2007T09)
A catalogue or marketplace listing
Prague Czech-English Dependency Treebank 2.0 (over two million words in close to 100,000 sentences)
Price
$100.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 15 June 2012.
Source
LDC catalog, Prague Czech-English Dependency Treebank 2.0 (LDC2012T08)
A catalogue or marketplace listing
Spanish Gigaword Second Edition (1.09 billion words, 3.7 million newswire documents)
Price
$4,000.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus released 17 July 2009.
Source
LDC catalog, Spanish Gigaword Second Edition (LDC2009T21)
A catalogue or marketplace listing
Treebank-2 (over 1.6 million words of hand-parsed Dow Jones and other text)
Price
$1,500.00 Non-Member
per corpus (research-only licence)
When
Oct 2026
What to know
Non-member fee: research use only, LDC bars non-members from commercial use; a for-profit LDC member pays the lower reduced-licence fee for commercial rights on top of membership. Corpus first released 1995 (catalog LDC95T7).
Source
LDC catalog, Treebank-2 (LDC95T7)
A catalogue or marketplace listing
5 million real consumer reviews and complaints with intent labels (PissedConsumer data)
Price
Average price per record/row is $0,15
per record (review)
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. The full 5M set lists at $736,800; final price depends on languages and licensing scope. Listed 7 Jul 2026.
Source
Opendatabay, 5M Consumer reviews & Complaints for NLP Training Data
A catalogue or marketplace listing
High-quality Arabic web corpus (FineWeb-edu Common Crawl text machine-translated by an LLM), 13.3M documents, 14.9B tokens
Price
$15,800
per corpus of 14,895,034,936 tokens
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. Derived from Common Crawl and translated by machine, so synthetic text under a custom licence; listed 17 Oct 2025.
Source
Opendatabay, High Quality Arabic Corpus (Token Haven)
A catalogue or marketplace listing
High-quality Spanish, Arabic and Norwegian web corpus (FineWeb-edu Common Crawl text machine-translated by an LLM), about 40M documents, 45B tokens
Price
$36,800
per corpus of 45B tokens
When
Oct 2026
What to know
Asking price on the Opendatabay marketplace (USD since 1 Oct 2026, converted from GBP at $1 = GBP 0.76); seller-stated volumes and rights, not verified by us. The bundle of the three single-language corpora; listed 20 Oct 2025.
Source
Opendatabay, High Quality Multilingual Corpus (Token Haven)
A catalogue or marketplace listing
GPU cost to machine-translate 13.4M FineWeb-edu web documents into Norwegian with an LLM (15.3B tokens)
Price
15,000 H100 Hours per language (~$36000)
per language (15,263,569,765 tokens)
When
Oct 2026
What to know
The producer's own stated compute cost; it sells the same corpus for $15,800, so synthetic translated text lists below its own production cost.
Source
Opendatabay, High Quality Norwegian Corpus (Token Haven)
A vendor's published price guide
Times of India and Economic Times content (Bennett, Coleman & Co) surfaced in ChatGPT under an OpenAI partnership
Price
around $5 million annually
per year
When
Sep 2026
What to know
Industry-source estimate; neither party disclosed terms. Shows what a large non-US news group gets versus $20M to $50M for top US brands.
Source
exchange4media
Press report
Indian Express Group content licensed to OpenAI for ChatGPT
Price
about $3 million a year
per year
When
Sep 2026
What to know
Industry-source estimate; terms undisclosed by the parties.
Source
exchange4media
Press report
News analysts, reporters and journalists evaluating AI-generated news content (4+ years' experience)
Price
$70-$110 per hour
per hour
When
Aug 2026
What to know
What a working journalist is paid by the hour to grade AI news output, against nothing per article when a publisher licenses their archive.
Source
Mercor job board, News analysts, reporters and journalists evaluating AI-generated news content
The payer's own page
Reddit other revenue (mostly data licensing, plus products sold directly to users), second quarter of 2026, all licensees
Price
$43 million, up 24% year over year
per quarter (all licensees)
When
Jul 2026
What to know
Aggregate, not one deal: Reddit's 10-Q says substantially all its licensing contract value comes from two partners, unnamed. $43 million of $805 million total revenue is about 5% (our arithmetic; the 10-Q gives $43,280 thousand of $804,905 thousand).
Source
Reddit Q2 2026 results, Exhibit 99.1 (SEC)
The payer's own page
Total damages awarded on appeal to 78 rights holders of Communication Books titles copied into the National Institute of Korean Language written corpus (Woongjin and Woongjin Booxen), 2026
Price
약 344만 원 (about 3.44 million won)
total award (78 rights holders)
When
Jul 2026
What to know
About 62 percent of the roughly 5.5 million won claimed. Separate from the 2023 settlement between the Korean Publishers Association, the National Institute of Korean Language and Woongjin Booxen, whose amount was not disclosed.
Source
Bizhankook, Woongjin loses AI copyright lawsuit (5 August 2026)
Press report
Per-title damages the Seoul Central District Court (appeal) set for e-books copied without permission into the National Institute of Korean Language written corpus, 78 rights holders of Communication Books titles against Woongjin and Woongjin Booxen, 2026
Price
전자책 정가의 70%에 인쇄 부수 5부 (70% of the e-book list price, times 5 copies, times the plaintiff's share)
per title, as a multiple of the e-book list price
When
Jul 2026
What to know
A court-set price for unlicensed use of a book in a training corpus: 3.5 times the e-book list price per title, times each plaintiff's share; the court held the use was not fair use. The appeal upheld a first-instance ruling of February 2025. The judgment itself was not read.
Source
Bizhankook, Woongjin loses AI copyright lawsuit (5 August 2026)
Press report
Bartz v. Anthropic class settlement: the $1.5 billion fund divided over the 482,460 works on the final Works List (books downloaded from pirate libraries)
Price
$1.5 billion for 482,460 Works
per settlement covering 482,460 works
When
Jul 2026
What to know
Gross settlement value per listed work, from the final approval order; the court itself rounds it to 'approximately $3,000'. It pays for past acquisition of pirated copies (the court found training itself fair use), so it prices liability, not a training licence going forward.
Source
Bartz v. Anthropic, Order Granting Final Approval of Class Action Settlement (N.D. Cal., Dkt. 680, 20 July 2026)
Court record
Bartz v. Anthropic class settlement: the fund left after the court-awarded fees, expenses, cost reserve and service awards, divided over the 440,490 works claimed by April 2026
Price
$1.5 billion less $101,561,111 fees, $2,635,197.46 expenses, $18,220,000 cost reserve and $45,000 service awards, over 440,490 claimed works
net settlement fund for 440,490 claimed works
When
Jul 2026
What to know
Because the fund is non-reversionary and paid pro rata per claimed work, the 8.7 percent of works never claimed raise the share of those claimed: net of the court's cut of class counsel to about 6.8 percent, a claimed work carries about $3,127 before the author and publisher split it (by default 50-50 for trade books). First payments were expected on or before 15 November 2026.
Source
Bartz v. Anthropic, Order Granting Final Approval of Class Action Settlement (N.D. Cal., Dkt. 680, 20 July 2026)
Court record
Books in the Bartz v. Anthropic class settlement (pirated copies used for training)
Price
approximately $3,000 per work, from a $1.5 billion fund
per book
When
Jul 2026
What to know
A litigation settlement for past unlicensed use, not a licence price; before fees, split between author and publisher.
Source
Authors Guild (on the court's final approval)
Court record
Wiley scholarly and learning content licensed for AI (all licensees, fiscal year to 30 April 2026)
Price
$49.1 million (prior year $40 million)
per year (all AI licensees)
When
Apr 2026
What to know
Aggregate revenue, not one deal; $19.4M of it came from other publishers' content Wiley resold for a royalty.
Source
John Wiley & Sons Form 10-K (SEC)
The payer's own page
News Corp US and UK news content, including archives, licensed to Meta for AI
Price
up to US$50 million a year, at least three years
per year
When
Mar 2026
What to know
Terms from people familiar; News Corp confirmed the deal but not the money.
Source
Mediaweek (citing The Wall Street Journal)
Press report
What one academic author received as his share when a small, ethical academic publisher licensed one of his monographs to an unnamed AI training company (opt-in by tick box, publisher keeps 50% of the total fee, buyer undisclosed under NDA)
Price
£500
per title (author's cheque after the publisher's 50% cut, one-off)
When
Feb 2026
What to know
First-hand account by the payee of a payment received at the end of 2025. The publisher and the AI company are not named (the publisher was under an NDA), so this is one self-reported data point, but it is an opt-in programme at a 50/50 split, like HarperCollins. The author calls the amount 'rather pathetically small' and the most he has made from writing in a year.
Source
Richard Pinner, Uniliterate, Academic Publishing AI Deals: My Biggest Paycheque Ever (24 February 2026)
The payer's own page
Author's default half of the per-work award in the Bartz v. Anthropic settlement (trade and university press books on the Works List)
Price
approximately $3,000 per work, by default split 50-50 between authors and publishers
per work (author's default 50% share)
When
Jan 2026
What to know
Damages for past use of pirated copies, not a licence price; gross of fees (the court later cut class counsel's award to about 6.8% of the fund). Education works have no default split. The gross $3,000 is on the board as bartz-anthropic-per-work.
Source
Bartz v. Anthropic, court-approved long-form class notice (updated deadline 8 Jan 2026)
Court record
High end of the per-book AI licensing offers Created by Humans says it has seen for authors' books (founder's statement, no buyer named)
Price
tens of thousands per book
per book
When
Nov 2025
What to know
An intermediary's own claim about offers it has received, with no buyer, title or deal named, so it is an upper-end signal only. Created by Humans says it takes a 20% fee and declines 'low ball offers' without naming their level.
Source
Written Word Media podcast, AI Rights and the Future of Publishing with Created by Humans founder Trip Adler
Press report
AI training royalty per title paid to History Press (Arcadia Publishing) authors who stayed in Arcadia's 2025 sublicence to an unnamed technology company; one author took it for three titles
Price
$1,020, or $340 per book
per title (author's royalty, one-time use)
When
Nov 2025
What to know
Author-side money in a sublicence that went ahead (inclusion was the default and this author accepted for all three titles). Two other authors were offered the same $340; Arcadia's own fee and the buyer are undisclosed. Arcadia has a catalogue of nearly 17,000 titles.
Source
Slate, Arcadia Publishing: Your town's local history books might be used to train A.I.
Press report
One-time AI training royalty Arcadia Publishing offered a Belt Publishing author (Rethinking Fandom, 2022) in its second round of sublicensing to an unnamed major technology company, opt-out by default
Price
$355
per title (author's royalty, one-time use)
When
Oct 2025
What to know
Arcadia's email reprinted in full by the author, who has one book with Belt and opted out. Opt-out deadline 7 November 2025, payment expected in spring 2026. Belt's director said the second-round titles were finalized in late October 2025.
Source
Cup of Coffee newsletter (Craig Calcaterra), 3 November 2025, reprint of Arcadia Publishing's 'Royalty Opportunity' email
The payer's own page
Bartz v. Anthropic class settlement: the payment Anthropic owes for each class work added to the Works List above 500,000
Price
$3,000/book for each Class work above 500,000
per additional book
When
Sep 2025
What to know
The settlement's marginal price per book, written into the agreement: the fund of $1.5 billion rises by $3,000 for each work above 500,000. The final Works List stopped at 482,460, so the clause was not triggered; it is the per-work rate the parties agreed, not a payment made.
Source
Bartz v. Anthropic, Class Action Settlement Agreement (N.D. Cal. 3:24-cv-05417, Dkt. 363-3), on the settlement website
Court record
Publisher content used in Perplexity's Comet browser and assistant (revenue share)
Price
80% of $5 monthly Comet Plus revenue; $42.5 million pool for early partners
total pool (plus 80% of subscription revenue)
When
Aug 2025
What to know
A usage-based revenue share, not a flat licence; Perplexity keeps 20% to cover compute.
Source
Axios
Press report
Royalty Bloomsbury offers academic and professional authors who opt in to AI training licences of their backlist titles
Price
a royalty rate of 20 percent
share of licence revenue per title
When
Aug 2025
What to know
A split, not a dollar figure: the publisher keeps 80%, against 50% at HarperCollins. Bloomsbury's 2025/26 interim results report its first non-exclusive AI licensing agreement without a fee.
Source
Publishers Lunch, Bloomsbury explores AI licensing
Press report
New York Times news, cooking and Athletic content licensed to Amazon for AI training and summaries
Price
between $20 million to $25 million per year
per year
When
Jul 2025
What to know
First AI licence for the Times; figure from people familiar, not confirmed by either company.
Source
TheWrap (citing The Wall Street Journal)
Press report
Author's expected return when Johns Hopkins University Press licenses a backlist title to an unnamed LLM developer (opt-out programme, about 3,000 backlist titles)
Price
less than $100 per title per license
per title, per licence (author's share)
When
Jul 2025
What to know
The press director's own guidance to authors; the buyer and the press's total fee were not disclosed. An upper bound only, and about 1/25th of the HarperCollins author share.
Source
The Baltimore Banner, Johns Hopkins Press will license its books to train AI
Press report
Taylor & Francis (Informa) incremental non-recurring AI data licensing revenue in the first half of 2025, including a new 2025 contract with a different, undisclosed LLM customer (academic books and content)
Price
about a £15 million incremental, non-recurring data licensing revenue number in the first half
revenue in the half year to 30 June 2025
When
Jul 2025
What to know
The seller's own finance director on its results call. Informa would not name the customer or the terms ('fairly under wraps'); its 2025 annual report adds only that it signed a further data access contract with another AI technology provider and that non-recurring data revenue was lower than 2024's $75m+.
Source
Informa plc, Half Year Results 2025, transcript of the results call (23 July 2025)
The payer's own page
One-time AI training royalty Arcadia Publishing offered the author of a local-history title (Images of Baseball: Kane County Cougars) for a sublicence to an unnamed major technology company, first round, opt-out by default
Price
$205
per title (author's royalty, one-time use)
When
Jul 2025
What to know
Arcadia's own email, posted as a screenshot by the author; the buyer and Arcadia's own fee were not disclosed. Inclusion was the default ('No action is needed to be included'), opt-out by 18 July 2025. Slate (3 November 2025) reports the same round paid other History Press authors $340 a title.
Source
Bluesky, David Malamut's post of Arcadia Publishing's email to authors (screenshot)
The payer's own page
What one Taylor & Francis author says they earned for their book under the publisher's AI training licence to Microsoft
Price
only $97
per book (author's royalty)
When
Jun 2025
What to know
One anonymous author's receipt, quoted by Bloomberg (paywalled) and repeated by Slate; Taylor & Francis pays authors under their existing contract royalty terms and gave no opt-out. Close to the under-$100 per title Johns Hopkins University Press told its authors to expect.
Source
Slate (3 November 2025), citing Bloomberg, Academic publishers sign AI deals as Trump cuts research funding (10 June 2025)
Press report
HarperCollins AI training licence fee actually paid to one author for one nonfiction backlist title (Writing Without Bullshit, 2016): three-year opt-in licence, author signed and received a cheque about a month later
Price
$2,500
per title (author's share of the per-title fee, before the agent's 15%)
When
Apr 2025
What to know
First-hand account by the payee that the money was paid: he signed and a month or so later received the cheque (shown as an image on the page). It confirms one author was paid, not how many authors opted in or were paid. The buyer is not named by HarperCollins; Bernoff calls it 'this AI company (likely Microsoft)', so Microsoft is his inference. The programme is the one already held as an offer (harpercollins-microsoft-author-share-per-title) and as the gross (book-title); this is the realised payment.
Source
Josh Bernoff, Why I took the $2,500 check to license my book for AI (bernoff.com; withoutbullshit.com redirects to it)
The payer's own page
Black Inc (Melbourne trade publisher): contract variation asking its authors to grant AI training rights to their books, author's share of the licence receipts
Price
net receipts split with the author 50/50
share of net licence receipts per title
When
Mar 2025
What to know
The publisher's offer to its own authors, not a buyer's price: no licensee or fee per title was disclosed. The Australian Society of Authors called 50/50 unfair and backs 75 percent to the author, as the Authors Guild does.
Source
The Guardian (Australia): Black Inc asks writers to sign AI agreements
Press report
Author's half of HarperCollins' per-title AI training fee (select nonfiction backlist, three-year licence, buyer reported as Microsoft)
Price
$2,500 per book
per book (author's share of $5,000)
When
Feb 2025
What to know
What the author receives: the publisher keeps the other $2,500. Opt-in; outputs capped at 200 consecutive words or 5% of a book. The $5,000 gross is already on the board as book-title.
Source
The Decoder, Microsoft offers authors $5,000 to train AI on their books
Press report
Royalty Cambridge University Press pays authors who opt in when their work is used in a generative AI licensing deal
Price
a royalty of 20%
share of licence revenue per title
When
Feb 2025
What to know
A split, not a dollar figure; per the post, licence revenue is divided equally across all content items in the licensed collection before the 20% applies.
Source
Silicon Reckoner (Michael Harris), quoting Cambridge University Press's letter to authors
Press report
Share of AI licensing income that some academic presses offer their authors per title, as reported by the Authors Guild (presses not named)
Price
a mere 25 percent
share of AI licence income per title
When
Dec 2024
What to know
A split, not a dollar figure, and the presses are unnamed. It sits between Bloomsbury and Cambridge (20%) and HarperCollins and Black Inc (50%); the Guild calls 75 to 85 percent to the author fair.
Source
Authors Guild, AI Licensing for Authors: Who Owns the Rights and What's a Fair Split?
Press report
One nonfiction book title, licensed for training for three years (HarperCollins)
Price
$5,000 per title
per title
When
Nov 2024
What to know
A signed deal between HarperCollins and an AI company reported as Microsoft, paid for each backlist title whose author opts in, split equally between author and publisher. Agents called it a one-off done by contract addenda.
Source
Publishers Weekly
Press report
Wiley's learning and research content licensed to tech companies for AI, total earned to date by August 2024
Price
$23 million
earned to date from AI deals
When
Aug 2024
What to know
Wiley's first AI licensing money, with a further $21 million then expected in the same financial year; authors were given no opt-out. Wiley's later fiscal years are on the board as wiley-fy2026-ai-licensing.
Source
infoDOCKET, citing The Bookseller
Press report
Dotdash Meredith (now People Inc.) magazine and web content licensed to OpenAI (2024 deal)
Price
reported around 16 million dollars a year
per year
When
May 2024
What to know
A reported figure; the publisher has described the deal as a guaranteed fixed payment plus a variable part tied to display in ChatGPT.
Source
LLM Pulse, AI content licensing deals 2023 to 2026
Press report
News Corp's current and archived journalism licensed to OpenAI for five years (2024 deal)
Price
reported 250 million dollars over five years
over five years
When
May 2024
What to know
The largest known publisher deal, reported and not confirmed by the parties; it covers training and display in ChatGPT. The per-year figure is the total divided evenly by the term.
Source
LLM Pulse, AI content licensing deals 2023 to 2026
Press report
Reddit's user posts licensed to OpenAI for training and display (2024 deal)
Price
reported around 70 million a year
per year
When
May 2024
What to know
A reported figure the parties have not confirmed; the deal also covers real-time display in ChatGPT, so not all of it pays for training.
Source
LLM Pulse, AI content licensing deals 2023 to 2026
Press report
Taylor & Francis academic books and journals licensed to Microsoft for AI (first year of the deal)
Price
almost £8m ($10m) in its first year
first year of the deal
When
May 2024
What to know
The $10m is the source's own conversion of about £8m. Authors were not told, could not opt out and were paid nothing extra, which is the rights risk a buyer prices in.
Source
ISSN International Centre newsletter, Taylor & Francis sells access to research to Microsoft AI
Press report
General text sold through a data marketplace (Defined.ai's quoted rate)
Price
$0.001 a word
per word
When
Apr 2024
What to know
$1,000 per million words. Edited, published work selected by a buyer has licensed for fifty times that.
Source
Reuters, via The Decoder
Press report
Financial Times journalism licensed to OpenAI (2024 deal)
Price
reported 5 to 10 million dollars a year
per year
When
Apr 2024
What to know
A reported range the parties have not confirmed. Covers training plus display with attribution in ChatGPT.
Source
LLM Pulse, AI content licensing deals 2023 to 2026
Press report
Reddit's user posts licensed to Google for AI training and access (2024 deal)
Price
a reported 60 million a year
per year
When
Feb 2024
What to know
A reported figure the parties have not confirmed. Reddit's own IPO filing disclosed $203.0 million of aggregate contract value across its January 2024 data licences, over two to three years.
Source
LLM Pulse, AI content licensing deals 2023 to 2026
Press report
Reddit's data licences signed in January 2024 (access to search and analyse Reddit data), aggregate of several contracts
Price
$203.0 million aggregate contract value, terms of two to three years
aggregate contract value of several licences, two to three years each
When
Jan 2024
What to know
The seller's own filing, the firmest figure behind the reported Google and OpenAI deals. It sums several contracts; Reddit expected at least $66.4 million of it as revenue in 2024.
Source
Reddit, Form S-1 registration statement (SEC)
The payer's own page
Apple's reported multiyear offers to news publishers (Conde Nast, NBC News, IAC) for archives to train its generative AI
Price
Reported offers of 50M-plus per publisher
per publisher (multiyear, total)
When
Dec 2023
What to know
Offers in talks first reported by The New York Times in December 2023; no signed deal has been confirmed, so this is the buyer's opening offer, not a licence.
Source
LLM Pulse, AI content licensing deals map (updated 28 Sep 2026)
Press report
Axel Springer news content licensed to OpenAI (December 2023 deal)
Price
a reported figure in the tens of millions of euros
per deal (term not stated)
When
Dec 2023
What to know
OpenAI's first global news publisher deal. Words only: the source gives an order of size, not a number, and does not say whether it is annual.
Source
LLM Pulse, AI content licensing deals 2023 to 2026
Press report
Per-title settlement Woongjin Booxen set with the publishers whose Bookopia e-books it used without permission in the National Institute of Korean Language written corpus (an AI and language-research corpus), Korea, 2022
Price
전자책 정가의 70%의 3copy (70% of the e-book list price, times 3 copies)
per title, as a multiple of the e-book list price
When
Sep 2022
What to know
A formula, not an amount: about 2.1 times the e-book list price per title. Woongjin called it a settlement of fees, not compensation. Seoul Shinmun (26 August 2022) reports 6,194 of the 15,993 disputed titles were settled at the level paid to publishers under formal contracts.
Source
Kyunghyang Shinmun, opinion column by culture researcher Lee Yung-hee, 13 September 2022
Press report
Korean government (National Institute of Korean Language) written-corpus collection project: total project cost over the 20,053 published works the contractor used from 1,226 publishers, 2019
Price
총 사업비는 30억 9000만원 (3.09 billion won) for 2만 53종 (20,053 works)
per project of 20,053 works
When
May 2019
What to know
A buyer paying a vendor to assemble a corpus, so the cost covers processing as well as rights; only 4,060 works were licensed from their publishers, and 15,933 Bookopia e-books were used without consent (later settled, see the related Korean rows). Bizhankook reports the contract required at least 80% of the fee to go to copyright clearance.
Source
Seoul Shinmun, Woongjin Booxen infringed copyright in 16,000 titles (26 August 2022)
Press report

What yours might fetch

A first answer or two gives a range. The full valuation asks the rest and shows where every figure comes from.

What is your data worth?

Kind of data

Size

Kind of writing

$20,000to$200,000

On two answers, against 2 public prices. An indication, not an offer.

Full valuation

No data upload
No account
Free

Why labs pay for it

Labs have been taken to court over text they did not license, and one settled for about $3,000 a book. Edited, expert writing with a clear rights holder is now bought, not scraped.

SteadyLarge publishers' licences are holding (one reported $49M of AI revenue in its 2026 fiscal year). Smaller and non-English publishers are mostly offered revenue shares, not cash.

Who holds it

Publishers and authors
Newspapers and magazines
Journals and societies
Forums and communities

What it looks like

A publisher's backlist
A regional newspaper's archive
Trade and technical journals
Writing in languages the big datasets miss

Raises the price

Edited, expert or professional writing
Rights held cleanly by one party
A subject or a language models are weak in
Regular new material, not only an archive

Lowers it

Text already freely available online
Rights split between authors, agents and publishers
User posts whose authors never agreed to licensing

The rights question

A publisher often holds print rights without holding AI training rights; authors may need to agree title by title. User-written content belongs to its writers unless the terms say otherwise.

How we take it to buyers

We represent archives with one clear rights holder. Collections of small publishers are stronger sold together, and we say when a revenue-share scheme is the realistic route.

Who buys it

Protege

A revenue share per deal, by the agreed partnership terms.

Protege, data providers page

Defined.ai

No rate card. It states "$600K+ average partner revenue" and "$1M+" for five or more partners.

Defined.ai, partnerships page

Datarade

The seller sets the price. Datarade takes 30% commission, or 20% to 15% on a paid plan of $6,000 to $12,000 a year.

Datarade, provider page

Created by Humans

Individual deals expected to run from hundreds to thousands of dollars, in proportion to a book's sales (its chief executive, reported).

Poets & Writers

FileYield

No figures. It charges sellers 8% and buyers 6% commission.

FileYield, sell data page

Opendatabay

The seller keeps 70% of a sale up to $100, rising to 95% above $250,000 (the commission falls from 30% to 5% as the sale grows). In a brokered sale the seller gets the agreed provider price and Opendatabay sets its own price to the buyer.

Opendatabay, fees and commission schedule

Every buyer and what it publishes

Where text comes from

The whole supply map

3,215,245 submissions
arXiv has received 3,215,245 submissions since August 1991.
Reaches buyers: Open download (bulk access); each paper keeps its author's licence
arXiv, Monthly submissions statistics, 16 Oct 2026
440,490 of the 482,460 Works
The Bartz v. Anthropic settlement covers 482,460 works on its Works List. By April 16, 2026, claims had been filed for 440,490 of them, a 91.3% claims rate.
Reaches buyers: Litigation settlement
CourtListener RECAP, Dkt. 643 Plaintiffs' Updated Claims Report, 16 Apr 2026
more than four million; more than 3.5 million (self-published); 642,242 (traditional)
More than four million books with ISBNs were published in the US in 2025. More than 3.5 million were self-published, and 642,242 came from traditional publishers.
Reaches buyers: Direct licence from publishers and authors, or litigation
Publishers Weekly, Book Output Topped Four Million in 2025, 17 Mar 2026
500,000:1; ~25,000:1 to ~100,000:1; 3,700:1
In 2025, Anthropic's crawlers fetched up to 500,000 pages for every visitor they referred back, settling at about 25,000 to 100,000 per referral after May. OpenAI's ratio peaked at 3,700:1 in March.
Reaches buyers: Crawl, increasingly blocked or metered at the CDN
Cloudflare Blog, The 2025 Cloudflare Radar Year in Review, 15 Dec 2025
2.17 billion web pages (or 361.4 TiB of uncompressed content); 587.2 million new URLs
The September 2026 Common Crawl archive (CC-MAIN-2026-39) holds 2.17 billion web pages (361.4 TiB uncompressed) from 40.4 million hosts. Only 587.2 million of its URLs had never been crawled before.
Reaches buyers: Open download (free public bucket)
Common Crawl, September 2026 Crawl Archive Now Available, 19 Sep 2026
~5%+ of all tokens in C4, or 28%+ of the most actively maintained, critical sources; 45% of C4 (ToS)
Between 2023 and 2024, robots.txt restrictions cut off about 5% or more of all tokens in C4 and 28% or more of its most actively maintained sources, and 45% of C4 was restricted by terms of service.
Reaches buyers: Crawl, now gated by robots.txt and terms of service
arXiv, Consent in Crisis: The Rapid Decline of the AI Data Commons, 20 Jul 2024

What a buyer's counsel will ask

23 questions apply to text, and 13 of them block a sale until settled. Most data that fails to sell fails here, not on price.

Run the rights check Read every question

Does the company selling the data own it, and has someone with authority approved the sale?
Blocks a sale until settled
How did the data leave each software system: a built-in export run by your own admin, or an app or script using the vendor's API?
Blocks a sale until settled
Does any of the data belong to or describe your customers, and do your customer contracts and data processing agreements allow using it for anything other than serving them?
Blocks a sale until settled
Does the data contain other companies' confidential information, such as partner documents, pricing, roadmaps or anything shared under an NDA?
Blocks a sale until settled
Does the data include personal information of employees, contractors or job applicants in California, and have they been told about the sale and given a way to opt out?
Blocks a sale until settled
For staff in the EU or UK, what legal basis lets you reuse the personal data in their work (messages, documents, code history, recordings) for a sale, and is that new purpose compatible with why you collected it?
Blocks a sale until settled

Other kinds of data

Company records Source code Video and footage Audio and speech Images Datasets, tasks and environments Physical-world data Expert and human work

Hold text? Describe it and we take it to the buyers.

Offer text I need this data

In this topic

Text and archives

The market rate for text, by side

What text has sold or been offered for

What yours might fetch

Why labs pay for it

Who buys it

Where text comes from

What a buyer's counsel will ask

Hold text? Describe it and we take it to the buyers.

What sells