FIND REMOTE AI JOBS
← All guides

What AI training jobs pay: rates from current listings

Advertised hourly rates for remote AI training work, by field and by platform, calculated from the roles listed in this directory.

Updated · 5 min read

This page reports advertised pay for remote AI training work. The figures are calculated from the 1,002 roles listed in this directory on 20 November 2026, of which 971 state an hourly rate in US dollars.

The short version

  • The middle advertised ceiling across all hourly roles is $70 an hour.
  • Among roles that also state a minimum, the middle starting rate is $55 an hour.
  • The highest advertised hourly rate is $400.

How rates are spread

Counting each role by its advertised ceiling:

Advertised ceilingRolesShare
Under $25 an hour919%
$25 to $49 an hour19020%
$50 to $99 an hour32834%
$100 an hour or more36237%

Pay by field

The middle figure is the median: half the roles in that field advertise more, half advertise less.

FieldRolesMiddle ceilingHighest
Health & law139$150 / hr$400 / hr
Business & finance96$100 / hr$350 / hr
Coding & AI116$90 / hr$300 / hr
Science & engineering170$88 / hr$200 / hr
Generalist453$51 / hr$200 / hr
Writing & creative149$40 / hr$150 / hr
Languages82$35 / hr$150 / hr

Pay by platform

PlatformRolesMiddle ceilingHighest
Mercor369$90 / hr$350 / hr
micro1358$85 / hr$400 / hr
SME by SuperAnnotate272$50 / hr$150 / hr
DataAnnotation3$125 / hr$150 / hr

How to read these numbers

  • A ceiling is not an offer. "Up to" rates are the most a role pays. Your rate depends on your assessed skills, your location, and the project.
  • Hours are not guaranteed. A high hourly rate on a project with few hours can pay less than a modest rate with steady work.
  • Rates differ by country. Some platforms publish different rates for the same role in different regions.
  • Per-task pay is separate. Roles paid per task, per month, or per year are left out of the hourly figures above.

To see roles at a given level, use the minimum hourly rate filter on the job list.