Skip to content
HumanSourcer

AI-training market data

Last updated

8160 active listings across 36 AI-training networks, rebuilt on every daily deploy. Every rate figure here is computed over 752 distinct advertised rates, not per listing - providers repeat one pay sentence across a whole batch, so counting listings would let a single advertisement vote thousands of times.

For the median rate itself and how it is derived, see the Pay Index.

Share this findingShare this findingShare this finding

How much do advertised rates vary?

The middle half of advertised AI-training rates falls between $25–$100/hr, but the full range runs from $1 to $400 per hour, with 8.5% of distinct advertised rates below $10/hr and 11.0% at $150/hr or above. Measured over 752 distinct advertised rates, not per listing.

$25–$100/hr · n=752

Share this findingXRedditLinkedIn

The rates split by how they were derived. 597 were stated outright, with a median of $65/hr; 155 came from text that hedged (“approximately”, “on average”), with a median of $25/hr. The two differ by more than double, so the cheap end of the distribution is substantially an artefact of how the figure was derived rather than what the market pays.

Why one pay sentence covers a thousand listings

6,321 listings state an hourly rate, but they carry only 752 distinct advertised rates - each one repeated across 8.4 listings on average. Providers post a batch of roles quoting one pay sentence, so counting per listing would let a single sentence vote thousands of times. Every rate figure on this page is computed over the distinct rates instead.

8.4× · n=752

Share this findingXRedditLinkedIn

Those rates are also unusually round: 73.9% are exact multiples of $5. These are advertised prices rather than negotiated ones, and advertised prices cluster.

How concentrated is this market?

Across 36 platforms with live listings, the largest single source accounts for 41.0% of indexed listings and the top three for 59.5%, giving a Herfindahl–Hirschman Index of 0.198 - moderately concentrated. This describes the listings HumanSourcer indexes, not the AI-training market itself: a platform that publishes no scrapeable board is absent from it entirely.

0.198 · n=36

Share this findingXRedditLinkedIn

How is pay actually quoted?

90.4% of listings that disclose pay quote an hourly rate. The other 689 quote per task, per unit, per project or as a salary, and are excluded from every hourly figure on this site - there is no honest way to convert a per-task price into an hourly one without knowing how long the task takes.

90.4% · n=7,181

Share this findingXRedditLinkedIn

How often is pay disclosed?

88.0% of active listings state some pay signal - but that is listing-weighted, and one large source posting thousands of near-identical roles moves it (Prolific Participants alone is the largest). Counted per platform instead, 30 of 36 (83.3%) publish pay on at least one open role.

88.0% · n=8,160

Share this findingXRedditLinkedIn

Where do listings say applicants must be?

Of the 3,262 listings whose title names a country (40.0% of 8,160), the most frequently named are USA (35.1%), United Kingdom (25.1%), Canada (5.7%). The remaining 4,898 state no country in their title, which is reported as unknown rather than assumed to be worldwide.

USA · n=3,262

Share this findingXRedditLinkedIn

How long do listings stay open?

Of 1,257 listings first indexed between 2026-10-15 and 2026-10-22, 60% had closed within 7 days. Only listings whose full 7-day window has elapsed are counted, so a listing posted yesterday cannot drag the figure down; the earlier bound is where closures stop being visible in the published data, so the cohort is a window rather than everything older.

60% · n=1,257

Share this findingXRedditLinkedIn

Which platforms advertise what

Platforms with at least 50 open roles. A median is shown only where the platform advertises at least 5 distinct rates - below that a single listing would set the figure, so the row shows advertised figures instead. “Up to” amounts are ceilings, not minimums or typical earnings. An em-dash means the platform states no hourly rate.

PlatformOpen rolesMedianMiddle halfDistinct ratesPay disclosed
Mercor Experts233$100/hr$69–$12572100%
AfterQuery Experts229$93/hr$64–$12243100%
micro1 Experts Opportunities295$83/hr$50–$1209498%
Handshake AI242$83/hr$51–$11934100%
Alignerr540$75/hr$50–$12064100%
Sepal Expert Hub88$75/hr$60–$937698%
SME Careers264$51/hr$29–$9044100%
DataForce jobs124$50/hr$36–$54754%
Rise Data Labs116$48/hr$44–$831276%
Mindrift671$37/hr$22–$621695%
Prolific Participants3349$30/hr$20–$805794%
Meridial833$30/hr$13–$655999%
Turing AI Advancement Work238$30/hr$20–$1251116%
LXT AI54$24/hr$23–$271544%
TELUS International AI Community portal68$17/hr$6–$292293%
Data Annotation Contributor community51$12/hr$9–$202496%
Welo Data322$12/hr$9–$247442%
DataAnnotation112Up to $150/hr advertised, too few distinct rates for a median4100%
Share this tableXRedditLinkedIn

How many listings go up each day

Listings per day since 1 Aug, by use case and by domain. 1,122 new listings were first indexed in the 7 days to 28 Oct, up 10% on the 7 days before. The window opens on 1 Aug because the scraper fleet was still being built through 31 July, and the days before it measure that rollout rather than demand.

Coverage change, 24 Oct: Collection expanded to DataAnnotation and Remotasks boards and additional Surge Workforce and HFC opportunities. These listings were new to our collection; this does not establish when they were posted or an increase in hiring.

7,918 listings open on 28 Oct

Open listings per day by use case

Open is the number of listings still up at the end of each day. Past days are an upper bound: a listing that disappears and returns reads as continuously open.

01,0002,0003,00030 Aug11 Sep23 Sep4 Oct16 Oct28 Oct
Research / analysisAI model evaluationCoding / software engineeringTranslation / linguisticsVoice / audioAll other use cases
Show data table
Open listings per day by use case, one row per day
DayResearch / analysisAI model evaluationCoding / software engineeringTranslation / linguisticsVoice / audioAll other use cases
2026-10-281,3762,5117358211,0291,446
2026-10-271,3542,6077598261,0851,459
2026-10-261,2762,5917598081,1011,437
2026-10-251,2462,5947618131,0911,443
2026-10-241,2332,6007628111,0911,446
2026-10-231,1812,5787477671,0861,454
2026-10-221,1082,5847517581,0821,454
2026-10-211,5402,5757527651,0991,417
2026-10-201,5282,5817527701,1031,447
2026-10-191,5362,5817497801,1001,439
2026-10-181,5362,5877527811,0981,453
2026-10-171,5392,5857547821,0981,453
2026-10-161,6052,5677647871,0601,464
2026-10-151,5772,5547567831,0401,449
2026-10-141,1942,5367637521,0591,455
2026-10-139442,5267607481,0501,451
2026-10-129392,5247157621,0481,450
2026-10-119472,5157177711,0411,452
2026-10-109472,5127197731,0391,454
2026-10-099502,4847207841,0421,486
2026-10-089512,4897118371,0451,488
2026-10-079472,4917068411,0451,484
2026-10-069492,4846818411,0461,459
2026-10-059512,4846568451,0411,467
2026-10-049532,4916618441,1121,467
2026-10-038892,4615577799761,345
2026-10-028922,4725657849891,358
2026-10-018932,4975667869931,366
2026-09-308852,4205547649901,326
2026-09-298872,4225517649781,345
2026-09-288882,4225417589771,341
2026-09-278882,4225427589801,340
2026-09-268892,4235407609821,345
2026-09-258812,4225407579621,342
2026-09-248852,4305417619651,379
2026-09-238882,4265427659551,379
2026-09-228932,4215367679261,388
2026-09-218872,4335267679321,402
2026-09-208872,4345257689381,398
2026-09-198882,4355277709381,399
2026-09-188912,4875317789271,407
2026-09-178772,4985287899281,413
2026-09-168642,4965277449141,380
2026-09-158652,5035517359071,377
2026-09-148552,4955447368981,349
2026-09-138552,4955447158951,350
2026-09-128552,4965457158961,353
2026-09-118532,5025557098961,353
2026-09-108702,5885737118981,412
2026-09-098562,5715597118911,387
2026-09-088552,5885507148981,392
2026-09-078592,5895607279031,393
2026-09-068602,5895607289031,393
2026-09-058622,5985657309031,401
2026-09-048422,5255346558931,356
2026-09-038372,5015366558601,360
2026-09-028212,4895286558591,389
2026-09-018172,5145356608051,406
2026-08-318202,5185546888161,407
2026-08-308212,5465796898141,412

Open listings per day by domain

Domain comes from the job title or, where the title names none, from a sentence in the listing that a model had to quote and that shows the work is in that domain. A listing can name several, so these lines overlap and do not add up to the total. 37% of listings still name no domain and appear on none of them. Open is the number of listings still up at the end of each day. Past days are an upper bound: a listing that disappears and returns reads as continuously open.

05001,0001,5002,00030 Aug11 Sep23 Sep4 Oct16 Oct28 Oct
CodingScience / STEMVoice / SpeechLinguisticsHealthcareAll other domains
Show data table
Open listings per day by domain, one row per day
DayCodingScience / STEMVoice / SpeechLinguisticsHealthcareAll other domains
2026-10-285791,2209731,5505981,027
2026-10-276031,2711,0281,6016271,078
2026-10-265981,2751,0451,5866001,056
2026-10-255981,2781,0471,5916041,063
2026-10-246001,2801,0461,5896041,067
2026-10-235851,2631,0471,5435941,049
2026-10-225881,2621,0481,5345981,053
2026-10-215911,2621,0641,5475971,027
2026-10-205921,2631,0681,5595961,027
2026-10-195951,2741,0651,5695871,034
2026-10-185981,2761,0651,5725881,045
2026-10-176001,2781,0651,5725871,044
2026-10-166051,2821,0621,5765681,052
2026-10-155991,2571,0421,5765591,038
2026-10-145981,2621,0611,5565581,034
2026-10-135961,2551,0501,5425541,040
2026-10-125971,2041,0441,5575531,039
2026-10-115991,2071,0371,5605561,044
2026-10-106011,2081,0371,5615531,043
2026-10-096351,2091,0401,5625271,042
2026-10-086341,2051,0441,5635261,053
2026-10-076311,1981,0441,5735221,056
2026-10-066301,1711,0261,5765221,055
2026-10-056301,1551,0191,5795021,058
2026-10-046331,1551,0901,5785061,079
2026-10-035301,0679561,538485953
2026-10-025321,0699681,543487960
2026-10-015351,0749731,545487970
2026-09-305221,0579721,523474930
2026-09-295181,0519561,526468942
2026-09-285091,0519551,519465941
2026-09-275101,0509581,519465941
2026-09-265101,0459601,520466947
2026-09-255111,0339391,521466947
2026-09-245121,0359391,522467945
2026-09-235141,0359291,524468938
2026-09-225101,0339021,518469940
2026-09-215071,0259081,519472947
2026-09-205051,0279141,520472949
2026-09-195071,0289141,522471951
2026-09-185111,0279031,522473969
2026-09-175101,0269041,522474978
2026-09-165091,0168881,468470974
2026-09-155321,0198811,465472974
2026-09-145311,0068711,460469960
2026-09-135311,0068681,439469960
2026-09-125331,0068691,439469962
2026-09-115441,0078691,434470958
2026-09-105581,0438701,435477999
2026-09-095471,0338621,426472997
2026-09-085371,0368611,4284731,014
2026-09-075471,0388651,4334731,015
2026-09-065471,0398651,4334731,014
2026-09-055511,0448651,4334751,018
2026-09-045339938521,365464922
2026-09-035359947991,323458919
2026-09-025289877921,310462886
2026-09-015339887381,311463867
2026-08-315531,0097361,342466872
2026-08-305791,0067341,340466903

Explore the data

Filter by domain, use case, provider or date. Every figure recomputes against the filtered set, and the filter state is kept in the URL so a view can be shared.

Advertised rate distribution

Each bar counts distinct advertised rates, not listings - providers repeat one pay sentence across a whole batch, so counting listings would let a single sentence vote thousands of times. Grey shows rates whose source text hedged ("approximately $20/hr"); those run materially lower than stated rates, so the left tail is partly an artefact of how the figure was derived.

No listings match these filters.
Show data table
Advertised hourly rate distribution, counted over distinct advertised rates
Hourly rateDistinct ratesStatedHedged
$0–$10000
$10–$20000
$20–$30000
$30–$40000
$40–$60000
$60–$80000
$80–$100000
$100–$150000
$150–$200000
$200+000

Median and middle half by group

Median advertised rate with the middle half (25th–75th percentile) as the bar. Computed over distinct advertised rates; a group with fewer than 5 of them shows an em-dash rather than a figure a single listing could set. Click a row to filter.

No listings match these filters.

Pay bands by provider

Share of each provider's listings by published hourly rate, top 10 providers by volume. "Unspecified" means no hourly rate could be determined from the listing. Click a row to filter by that provider.

No listings match these filters.

Demand heatmap

Darker = more listings for that use case × provider combination. Click a cell to filter by both.

No listings match these filters.
Listing counts by use case and provider. Empty cells mean no listings for that combination under the current filters.

How these figures are derived

Listings are scraped daily from the public boards of 36 AI-training platforms. Pay is read from a structured field where one exists, parsed by rule where it does not, and otherwise extracted from the listing prose by a model that must quote the sentence it took the figure from.

Rates are deduplicated before any percentile is computed. The key is the provider plus the verbatim pay sentence, so a batch of listings sharing one advertisement counts once. This is the single most important thing to know about the numbers above: the per-listing figures are materially different and would overstate the sample by roughly seven times.

Only rates quoted per hour are comparable, so per-task, per-unit and salaried listings are excluded from every rate figure and counted separately under how pay is quoted.

No rate trend is published. The index begins at 2026-07-19, so there is not yet enough history for a pay trend that would mean anything, and pay figures are re-derived whenever the extractor improves. Listing volume is different: first-seen and last-seen dates are never rewritten, so the listings per day series is reconstructed from them for every listing including closed ones, from 2026-08-01. The closure figure above uses a 7-day horizon and counts only listings whose full window has elapsed.

Domains come from two sources, and the second is gated. A domain is taken from the job title where the title names one. Where it does not, a model reads the listing and may assign a domain only by quoting a sentence from the listing that describes the work itself, not a required background or a tool. It is told to return nothing when no domain fits, so a share of listings still has none and is left out of the domain views rather than guessed at. The domain pages under /roles/ still use titles only.

Concentration describes the listings indexed here, not the AI-training market as a whole - a platform that publishes no scrapeable board does not appear at all.

The aggregated figures on this page are also available as JSON. Reuse terms are the same as the Pay Index.

Common questions

How concentrated is the AI-training listing market?
Across 36 platforms with live listings, the largest single source accounts for 41.0% of indexed listings and the top three for 59.5%, giving a Herfindahl–Hirschman Index of 0.198 - moderately concentrated. This describes the listings HumanSourcer indexes, not the AI-training market itself: a platform that publishes no scrapeable board is absent from it entirely.
Why do so many AI-training listings quote the same pay rate?
6,321 listings state an hourly rate, but they carry only 752 distinct advertised rates - each one repeated across 8.4 listings on average. Providers post a batch of roles quoting one pay sentence, so counting per listing would let a single sentence vote thousands of times. Every rate figure on this page is computed over the distinct rates instead.
Do all AI-training jobs pay by the hour?
90.4% of listings that disclose pay quote an hourly rate. The other 689 quote per task, per unit, per project or as a salary, and are excluded from every hourly figure on this site - there is no honest way to convert a per-task price into an hourly one without knowing how long the task takes.
How often do AI-training platforms publish pay at all?
88.0% of active listings state some pay signal - but that is listing-weighted, and one large source posting thousands of near-identical roles moves it (Prolific Participants alone is the largest). Counted per platform instead, 30 of 36 (83.3%) publish pay on at least one open role.
How many new AI-training listings appear each day?
1,122 new listings were first indexed in the 7 days to 28 Oct, up 10% on the 7 days before. "New" means first seen by HumanSourcer, not the date the provider posted it, and the daily series starts on 2026-08-01.
Are advertised AI-training rates round numbers?
Yes - 73.9% of the 752 distinct advertised rates are exact multiples of $5. Rates are advertised, not negotiated, so they cluster on round figures rather than spreading evenly.