AI Trainer Salary: What the Work Pays Per Hour in 2026
Open AI trainer roles on OpenTrain list a median of $70 per hour and a typical range of $35 to $110. See pay by role type and experience level.
Open AI trainer roles on OpenTrain list a median rate of 35 and 3 to $400 at the extremes. There is no salary in the usual sense because almost every role is a part-time hourly contract, so what you earn in a week is your rate multiplied by the hours you choose to work. Specialist review in law, finance, and medicine pays the most. Transcription, translation, and data collection pay the least but hire the widest range of people.
The figures below come from every open role on OpenTrain on 17 September 2026. The live pay-rate index refreshes hourly and breaks the same numbers down by country.
Pay by role type
Role type explains more of the difference in pay than anything else. Roles that need a professional credential sit at the top. Roles anyone with strong reading and writing can do sit in the middle. Roles that need a native speaker or a quiet room and a microphone sit at the bottom.
Median hourly rate by role type
Median of the listed hourly rate for open roles in each category. Categories with fewer than five hourly roles are omitted.
Legal and finance
$120/hr
160 hourly roles
General annotation
$100/hr
60 hourly roles
Medical and health
$95/hr
80 hourly roles
Writing and editing
$80/hr
13 hourly roles
Model evaluation and RLHF
$70/hr
372 hourly roles
Coding and software
$70/hr
161 hourly roles
Image and video annotation
$70/hr
29 hourly roles
Audio, speech, transcription
$35/hr
140 hourly roles
Translation and localization
$30/hr
100 hourly roles
Data collection
$20/hr
47 hourly roles
Two things stand out. Legal and finance roles list a typical range of 150, so even the lower end beats the median for most other work. And model evaluation, the largest category, spans 105 for the middle half of roles, which means the same job family can pay very differently depending on the subject matter and the experience required.
How rates are distributed
Averages hide the shape of the market. About a third of hourly roles list 40.
Open hourly roles by listed rate
Count of open roles in each rate band. Roles without a listed hourly rate are excluded.
Under $20/hr
78 roles
$20 to $39/hr
246 roles
$40 to $69/hr
256 roles
$70 to $99/hr
193 roles
$100 to $149/hr
257 roles
$150/hr and above
141 roles
Pay by experience level
Experience level in AI training means AI training experience, not career experience. That produces a pattern that surprises people: entry-level roles list a higher median than intermediate roles, because many entry-level roles are specialist roles that need a law degree or a medical license rather than prior rating work.
Median hourly rate by experience level
Entry level means no prior AI training experience is required. Many entry-level roles still require a professional credential.
Entry level
$80/hr
887 hourly roles
Intermediate
$45/hr
189 hourly roles
Expert
$70/hr
95 hourly roles
Intermediate roles cluster in audio, translation, and annotation work where the requirement is a proven track record on similar tasks rather than a credential. Expert roles are mostly senior coding and research review.
Turning an hourly rate into weekly income
Because the work is part time, the useful question is what a realistic week looks like. At the $70 median:
| Hours per week | Weekly income at $70/hr |
|---|---|
| 10 | $700 |
| 20 | $1,400 |
| 30 | $2,100 |
These are arithmetic, not promises. Hours depend on how much work a project has open, and rates on any single role can sit far from the median. The listing you apply to is always the source of truth.
What moves the rate
- Credentials. A law, medical, finance, or engineering background is the single biggest lever.
- Programming depth. Senior coding review pays at the top of the coding range.
- Language. Fluency in a language with fewer available raters commands a premium over English-only work.
- Country. Roles restricted to a specific country often list rates aligned to that market.
- Task type. Writing and rewriting pay more than simple rating. Recording and transcription pay less.
How to earn at the top of the range
- Lead with your specialist credential on your profile, even if you have never done AI work.
- Apply to roles in your own field before general evaluation roles.
- Add every language you can work in professionally.
- Take the assessment seriously. Rates on a project are fixed, but access to the higher-paying projects depends on how you score.
See what is open now on the AI training jobs board, check the live pay-rate index, or start with what an AI trainer does if you are new to the work.
Frequently asked questions
AI trainer work is almost always hourly contract work rather than a salary. Open roles on OpenTrain list a median of $70 per hour, and half of all hourly roles sit between $35 and $110. Weekly income depends on the hours you choose to work.
Specialist review roles pay the most. Legal and finance roles list a median of $120 per hour, general annotation with domain expertise lists $100, and medical roles list $95. Coding and model evaluation roles list a median of $70.
Not necessarily. Entry level means no prior AI training experience, and many entry-level roles are specialist roles that require a degree. Entry-level roles on OpenTrain list a median of $80 per hour, higher than intermediate roles.
Every figure is calculated from the pay listed on open roles on OpenTrain. Nothing is estimated or self-reported. The live index at opentrain.ai/jobs/pay-rates/ refreshes every hour.