Skip to content

Research report ยท October 2026 edition

How Long AI Training Roles Stay Open

Half of the AI training roles listed from 3 Aug 2026 to 31 Oct 2026 came down within 25.3 days of being listed; Writing & Content turned over fastest (median 9.2 days), and counting only removed roles would put the median at 13 days.

Read the month's story: The AI Training Job Market in October 2026

Provisional figures. This month is complete, but late data can still arrive until 3 Nov 2026, when the edition is frozen for good.

Data window
3 Aug 2026 to 31 Oct 2026
90 days, UTC
Computed
1 Nov 2026, 01:11 UTC
Final once frozen
Source
The public job boards of the AI training platforms this site tracks, read since 3 Aug 2026 (every hour since 18 Oct 2026).
Method
Version 3. Definitions
CC BY 4.0, how to cite

Every figure on this page as CSV and JSON, sent to your work email.

median lifetime of a listing (Kaplan-Meier)
25 days
of roles still listed four weeks after going up
45%
median counting only removed roles (too short)
13 days
of open roles listed 25+ days (evergreen)
70%

Key findings

  1. 1

    Half of the 3,167 AI training roles listed from 3 Aug 2026 to 31 Oct 2026 came down within 25.3 days of being listed (Kaplan-Meier median), and 44.7% were still listed four weeks after going up.

    Figure 1
  2. 2

    Counting only the roles already removed puts the median at 13 days; counting the 46.0% still listed too, it is 25.3 days, 95.0% longer.

    Figure 4
  3. 3

    Writing & Content turns over fastest (half its roles came down within 9.2 days, 34.2% still listed after four weeks); Transcription, Audio & Speech stays listed longest, with 53.7% of its roles still up after four weeks.

    Figure 2
  4. 4

    Alignerr takes its roles down fastest (half its roles came down within 4.7 days, 15.9% still listed after four weeks) and Invisible keeps them longest (100.0% still listed after four weeks).

    Figure 3
  5. 5

    14.0% of the removed roles were followed within 3 days by a listing with the same title on the same platform, so part of what reads as turnover is re-posting.

    Figure 3
  6. 6

    3,048 of the 4,381 roles open on 1 Nov 2026 (69.6%) had been listed for 25 days or more, longer than even the longest-lived kind of listing is likely to stay up, the most in Math & STEM (84.7% of its open roles).

    Figure 5
  7. 7

    In the median complete week, 9.9% of the open roles came down (from 6.5% to 18.3% across 12 weeks); the highest was the week of 3 to 9 Aug.

    Figure 6

Figure 1

How many roles are still listed, week by week

What share of AI training roles is still listed one, two, four or eight weeks after it goes up?

Age of the listingStill listed
After 1 week
82.7%
After 2 weeks
62.7%
After 3 weeks
53.8%
After 4 weeks
44.7%
After 6 weeks
31.6%
After 8 weeks
25.7%

Sample. 3,167 roles listed from 3 Aug 2026 to 31 Oct 2026 with a known listing date, 1,709 of them removed by 1 Nov 2026; back catalogues left out.

Kaplan-Meier estimate: a role that has come down ends at its last sighting; a role still listed is counted for as long as it has been seen, then leaves the count without being treated as removed.

Age runs from the listing date the job board states (or the first day the role appeared in the feed, if earlier or not stated), in hours. A role first seen more than 2 days after that date surfaced in the feed late and is counted only from the age at which it was first seen; within 2 days it is the daily read catching up and the role counts from its listing day.

A share is read from the roles listed at least that long before 1 Nov 2026, with their count beside it; ages no role of the cohort has reached yet read n/a.

Source. AITraining.jobs Research, How Long AI Training Roles Stay Open, October 2026 edition.

Figure 2

How long roles stay listed, by kind of work

Which kinds of AI training work turn over fastest, and which stay listed longest?

Kind of workStill listed after 2 weeksAfter 4 weeksMedian lifetime (days)
All kinds of work62.7%
44.7%
25.3
Transcription, Audio & Speech63.7%
53.7%
58.0
Law, Medicine & Finance64.4%
51.5%
29.5
Robotics & Physical-AI Data55.7%
47.8%
25.0
Coding & Software Evaluation60.5%
46.7%
23.1
Math & STEM69.5%
44.8%
27.1
Agentic Tasks & RL Environments47.9%
44.6%
13.4
Research Studies & Surveys56.5%
44.5%
24.1
Multilingual & Localization60.4%
44.1%
25.5
Data Annotation & Labeling52.9%
39.8%
18.0
Writing & Content44.6%
34.2%
9.2
AI Evaluation & Rating (RLHF)65.5%
31.3%
18.5
Red-Teaming & AI Safety71.3%
18.9%
18.3

Sample. 3,167 roles listed from 3 Aug 2026 to 31 Oct 2026 with a known listing date, by the kind of work they are classified as.

Shares still listed and the median lifetime are Kaplan-Meier estimates, so roles still listed count for as long as they have been seen.

Median lifetime n/a: more than half of the group's roles were still listed at the oldest age it can be read at, so its median is not reached yet (it is longer than the time observed), or fewer than 3 of its roles were listed long enough ago to read it (a middle needs that many). A share reads n/a when none of the group's roles was listed that long ago.

Kinds of work whose roles were observed listed for fewer than 433 days in all are marked as a small sample and not ranked: below that the pace of removal of the kinds of work together says more about them than their own removals (the weight of that pace, fitted on their own removals).

Source. AITraining.jobs Research, How Long AI Training Roles Stay Open, October 2026 edition.

Figure 3

How long roles stay listed, by platform

On which platforms do AI training roles come down fastest?

PlatformStill listed after 2 weeksAfter 4 weeksMedian lifetime (days)
All platforms62.7%
44.7%
25.3
micro174.9%
57.2%
31.5
Mercor62.1%
33.3%
21.0
OpenTrain AI52.8%
22.6%
17.9
Mindrift9.5%
2.3%
7.3
AfterQuery100.0%
99.3%
n/a
Alignerr15.9%
15.9%
4.7
Braintrust62.1%
47.7%
27.0
Turing94.2%
94.2%
30.0
TransPerfect DataForce74.1%
47.0%
27.8
Terac70.2%
52.2%
28.3
Welo Data36.5%
29.7%
9.8
Appen86.9%
48.3%
27.9
xAI (SpaceXAI)77.5%
54.3%
32.1
Sigma AI78.3%
67.3%
34.7
Handshake AI95.5%
95.5%
n/a
Volga Partners66.7%
n/a
n/a
Sovrano AI100.0%
n/a
n/a
Weekdayn/a
n/a
n/a
RWS TrainAI Communityn/a
n/a
n/a
SuperAnnotate (SME)51.3%
n/a
n/a
Anyone AI83.3%
n/a
n/a
Outlier (Scale AI)75.0%
50.0%
16.3
Invisible100.0%
100.0%
n/a
Lifted (Upwork)n/a
n/a
n/a
Babel AudioSmall samplen/a
n/a
n/a
OneFormaSmall samplen/a
n/a
n/a
RemoExpertsSmall samplen/a
n/a
n/a
Rise Data LabsSmall samplen/a
n/a
n/a
Scale AI Human Frontier CollectiveSmall sample100.0%
n/a
n/a
TELUS Digital AISmall samplen/a
n/a
n/a
CohereSmall samplen/a
n/a
n/a
Pasiflora AISmall samplen/a
n/a
n/a
Perle LabsSmall sample100.0%
n/a
n/a
Pila8Small samplen/a
n/a
n/a
Shaip (Collaborate by Shaip)Small samplen/a
n/a
n/a

Sample. 3,167 roles listed from 3 Aug 2026 to 31 Oct 2026 with a known listing date, by the platform that lists them. Listings only.

Median lifetime n/a: more than half of the group's roles were still listed at the oldest age it can be read at, so its median is not reached yet (it is longer than the time observed), or fewer than 3 of its roles were listed long enough ago to read it (a middle needs that many). A share reads n/a when none of the group's roles was listed that long ago.

Re-listed: a removed role followed, within 3 days of its last sighting, by a listing with the same title on the same platform (the window is the longest gap that is more likely a platform posting the role again than a later opening with the same title: the gaps from a removal to the next same-title listing, fitted by kinds, put 13.8% of removed roles in a re-posting kind with a mean gap of 1.4 days; shown when the platform's removed roles read on their own: at least 2 removed roles, the weight fitted on the platforms' own counts). 14.0% of all removed roles in the cohort were. A platform that takes roles down and posts them again, for example weekly, has short lifetimes here although the opening itself may run for months; listings are counted as the feeds publish them, not merged further.

Some platforms post one listing per country or language, each with its own lifetime.

Platforms whose roles were observed listed for fewer than 39 days in all are marked as a small sample and not ranked: below that the pace of removal of the platforms together says more about them than their own removals (the weight of that pace, fitted on their own removals).

Source. AITraining.jobs Research, How Long AI Training Roles Stay Open, October 2026 edition.

Figure 4

Why the removed-only median is too short

How much does counting only the roles that have come down understate how long roles stay listed?

Kind of workStill listed on 1 Nov 2026Median days, removed roles onlyMedian lifetime, all roles (days)
All kinds of work46.0%13.0
25.3
Robotics & Physical-AI Data65.8%12.5
25.0
Agentic Tasks & RL Environments65.2%8.3
13.4
Transcription, Audio & Speech61.8%8.2
58.0
Law, Medicine & Finance52.5%9.7
29.5
Research Studies & Surveys51.7%10.3
24.1
Math & STEM48.9%20.8
27.1
Data Annotation & Labeling44.1%11.5
18.0
Multilingual & Localization39.8%16.3
25.5
Coding & Software Evaluation37.0%13.1
23.1
Red-Teaming & AI Safety21.9%20.5
18.3
AI Evaluation & Rating (RLHF)20.4%18.0
18.5
Writing & Content18.5%8.8
9.2

Sample. The same 3,167 roles listed from 3 Aug 2026 to 31 Oct 2026; the kinds of work whose lifetimes read on their own (as in the figure by kind of work).

Removed roles only: the median days from listing to last sighting among the roles that have already come down, the measure a 'median days listed' in a hiring report states. It cannot see the roles still listed, which are the longer-lived ones, so it comes out short, the more so the more roles are still up.

All roles: the Kaplan-Meier median, which counts every role still listed for as long as it has been seen.

Median lifetime n/a: more than half of the group's roles were still listed at the oldest age it can be read at, so its median is not reached yet (it is longer than the time observed), or fewer than 3 of its roles were listed long enough ago to read it (a middle needs that many). A share reads n/a when none of the group's roles was listed that long ago.

A removed-only median needs at least 3 removed roles: a middle, the one minimum of every figure over independent units.

Source. AITraining.jobs Research, How Long AI Training Roles Stay Open, October 2026 edition.

Figure 5

Evergreen roles: open longer than any kind of listing lasts

How many of the roles open now have been listed for 25 days or more?

Kind of workListed 25+ daysShare of open roles
All kinds of work3,048
69.6%
Math & STEM210
84.7%
Multilingual & Localization742
84.5%
Data Annotation & Labeling183
76.9%
Coding & Software Evaluation223
73.8%
Transcription, Audio & Speech428
67.1%
Law, Medicine & Finance977
63.4%
Writing & Content19
57.6%
Robotics & Physical-AI Data57
57.0%
Red-Teaming & AI Safety30
55.6%
Research Studies & Surveys10
55.6%
AI Evaluation & Rating (RLHF)114
55.1%
Agentic Tasks & RL Environments55
44.4%

Sample. 4,381 roles open on 1 Nov 2026, of every listing date, back catalogues included.

Evergreen: open on 1 Nov 2026 and listed at least 25 days before it. That is the age at which even a listing of the longest-lived kind is more likely down than up: the lifetimes of the 3,167 roles listed from 3 Aug 2026 to 31 Oct 2026, fitted by kinds (1 kind), put 93.6% of them in the longest-lived kind, which stays up 36.3 days on average; a role open longer is kept open.

A back-catalogue role whose board states no listing date is aged from the first day this site saw it, so its age, and the count of evergreen roles, is a floor.

Rows with fewer than 14.6 open roles are marked as a small sample and not ranked: below that the average of the rows weighs more than a row's own open roles (the weight fitted on the rows' own counts).

Source. AITraining.jobs Research, How Long AI Training Roles Stay Open, October 2026 edition.

Figure 6

Weekly churn of open roles

What share of the open AI training roles comes down each week?

WeekShare removed in the weekOpen at the week's start
3 to 9 Aug
18.3%
1,483
10 to 16 Aug
9.2%
1,483
17 to 23 Aug
6.5%
1,560
24 to 30 Aug
11.2%
1,595
31 Aug to 6 Sep
9.7%
1,602
7 to 13 Sep
12.9%
2,232
14 to 20 Sep
7.4%
2,585
21 to 27 Sep
8.8%
2,584
28 Sep to 4 Oct
13.5%
2,663
5 to 11 Oct
8.5%
2,557
12 to 18 Oct
10.1%
2,800
19 to 25 Oct
10.6%
4,166

Sample. Every complete ISO week (Monday to Sunday, UTC) from 3 Aug 2026 to 31 Oct 2026, on the platforms already tracked at each week's start.

Churn: of the roles open at 00:00 UTC on the Monday, the share last seen during that week and gone from the feed by the end of the observation. Roles listed and removed within the same week are not in the rate; they are counted among the new listings.

Each week counts only platforms tracked before it began, so a platform added to the dataset never reads as churn or growth.

A platform that takes a batch of roles down at once (a project closing, a board reorganised) shows as a single high week.

Rows with fewer than 122.1 roles open at the week's start are marked as a small sample and not ranked: below that the average of the rows weighs more than a row's own roles open at the week's start (the weight fitted on the rows' own counts).

Source. AITraining.jobs Research, How Long AI Training Roles Stay Open, October 2026 edition.

For labs, data vendors and researchers

Buy the data behind this report

The figures above are a summary. Behind them are records we collect ourselves, every day:

  • 11,482 roles tracked, each with its history

    Every role on 62 job boards, read since 13 Jul 2026 (every hour since 18 Oct 2026): its pay and place, and the first and last day it was open.

  • 58,632 apply clicks in 28 days

    Which roles people actually apply to, by field, pay and kind of work. It happens on this site, so nobody else can count it.

  • 39 lab and vendor hiring boards

    The job boards of 25 AI labs and 14 human-data vendors, read every day: who is hiring for human data, and who is hiring to buy it.

What you can buy

  • The records

    The rows behind any figure, with their history, not only the totals.

  • A regular feed

    The same figures every week or month, as files.

  • A custom cut or benchmark

    Your countries, languages or fields, or your own platform against the market.

Building AI and need the experts behind this data? (you contract them yourself) or (we deliver the expert work).

Investing in AI-data companies? Hiring radar is these records role by role, every day, as files.

Method and definitions

The public job boards of the AI training platforms this site tracks, read since 3 Aug 2026 (every hour since 18 Oct 2026). Every figure is computed by one database query over the window shown, the same query for the live report and for every edition. Each figure states its sample beneath it.

Listing date
The date the platform's job board gives for the posting, or else the first day the role appeared in the feed. A role that arrived with a platform's back catalogue, or came into view with a change in how this site reads its board, and gives no date has no known listing date.
Cohort
Every role with a known listing date listed in the 26 weeks to the end of the period (from 3 Aug 2026, the start of this site's index, at the earliest), except each platform's back catalogue: the roles it already had up in its first week in this dataset.
Lifetime
From the listing date to the last day the role was seen in its platform's feed, in hours, stated in days. A role still listed at the end of the period has lasted at least that long: it is censored, not removed.
Still listed after N weeks
The Kaplan-Meier estimate of the share of roles still in their feed N weeks after their listing date. It counts every role for as long as it has been seen, so roles still listed are not mistaken for short-lived ones. A role first seen more than two days after its listing date surfaced late and counts only from the age it was first seen.
Median lifetime
The age by which half the roles are estimated to have come down. Not reached when more than half are still listed at the oldest age the data covers.
Removed-only median
The median days from listing to last sighting among roles that have already come down, the usual 'median days listed'. It leaves out the roles still up, which are the longer-lived ones, so it is shorter than the median lifetime.
Re-listed
A removed role followed by a listing with the same title on the same platform within the re-posting window: the gaps from a removal to the next same-title listing are fitted by kinds (most removed roles never come back), and a gap counts while it is more likely of the shortest kind, a platform posting the role again, than of every longer kind, a later opening with the same title.
Evergreen role
A role open at the end of the period and listed longer than the age at which even a listing of the longest-lived kind is more likely down than up: the cohort's lifetimes fitted by kinds, each with its own daily chance of coming down, and that age is the first whole day t with q to the power t at most one half, q the longest-lived kind's daily chance of staying up.
Weekly churn
Of the roles open at 00:00 UTC on a Monday, the share removed during that ISO week, on the platforms tracked before the week began.
Kind of work
Each role is classified from its title and description into one of the twelve kinds of AI training work this site tracks, from AI evaluation and coding to multilingual and audio work. Every kind links to its own page.
Window
Live: roles listed in the 26 weeks to yesterday (from 3 Aug 2026 at the earliest), in UTC. Edition: roles listed in the 26 weeks to the end of that month (from 3 Aug 2026 at the earliest), observed to the end of the month.

What this data cannot tell you

  • Platforms that recruit without a public job board are not counted. Several large ones invite workers to projects directly.
  • A listing is not a hire, and a removed listing may have been filled, paused or closed: platforms do not say which.
  • Some platforms take roles down and post them again (some weekly) or post one listing per country or language, so a short listing life is not always a short opening. The re-listing column shows where.
  • Feeds are read every hour (about once a day before 18 Oct 2026): a role listed and removed between two reads may be missed, and a removal is dated to the last read that still saw the role.
  • Ages beyond what the cohort has been observed for are not estimated, so a young edition shows the longer ages as n/a.
  • A platform that takes a batch of roles down at once (a project closing, a board reorganised) shows as a single high week of churn.

How the job feeds are read and roles are classified: site methodology. How fresh, complete and consistent the data is, checked every day: data quality and coverage.

Editions

The live report follows every role listed in the 26 weeks to the last complete day (from 3 Aug 2026 at the earliest), and changes every day. Each complete month is also published as an edition, frozen two days after the month ends, so a figure you cite from an edition never moves.

Cite and reuse

Free to quote, chart and republish under CC BY 4.0 with a link to this page. This edition is frozen, so its figures will read the same later.

Citation

AITraining.jobs (2026). How Long AI Training Roles Stay Open, October 2026 edition. https://aitraining.jobs/research/how-long-ai-training-roles-stay-open/2026-10

BibTeX

@misc{aitrainingjobs_how_long_ai_training_roles_stay_open_2026_10,
  author = {{AITraining.jobs}},
  title = {How Long AI Training Roles Stay Open, October 2026 edition},
  year = {2026},
  howpublished = {\url{https://aitraining.jobs/research/how-long-ai-training-roles-stay-open/2026-10}},
}
Data: every figure as CSV (one line per value) and JSON, sent to your work email.Pay and open-roles data

More research

All 81 research reports