Report
AI evaluator work: pay & coverage report
Summary
Specialist AI Work currently tracks 725 public-ready AI evaluation roles. 465 of 725 (64%) publish parseable hourly pay. The figures below are computed only from those roles and are not a market-wide wage survey.
Coverage by platform
Mercor190
Handshake19
Micro1317
Turing199
Coverage by profession
Software engineering158
Legal56
Medical44
Finance74
Writing and editing18
Language and bilingual196
Research and academia93
General professional86
Pay by profession
| Profession | Hourly pay (qualifying roles) |
|---|---|
| Software engineering | $6.5-135/hr (median $60, n=79) |
| Legal | $20-150/hr (median $70, n=45) |
| Medical | $20-175/hr (median $90, n=32) |
| Finance | $7-182/hr (median $80, n=51) |
| Writing and editing | $6-55/hr (median $32.5, n=12) |
| Language and bilingual | $12-182/hr (median $25.5, n=131) |
| Research and academia | $7.5-175/hr (median $73.5, n=51) |
| General professional | $2.5-165/hr (median $56.3, n=64) |
Advertised pay ceilings
19roles advertise pay as a ceiling ("up to $X/hour"). Advertised ceilings range $75-175/hr (median $80). A ceiling is the maximum advertised rate, not a typical or guaranteed rate, so it is reported separately from firm hourly pay.
Eligibility
721 of 725 tracked roles are remote. 54 mention U.S. location or work-authorization cautions.
Methodology & limitations
- - Source is a single independent tracker (Specialist AI Work), not a market-wide wage survey.
- - Pay is self-reported by platforms and parsed conservatively; vague, non-hourly, referral-bonus, and suspicious pay are excluded.
- - Some platforms advertise pay as a ceiling ("up to $X/hour"); these are reported separately as advertised ceilings, not mixed into firm hourly rates.
- - Statistics exclude held, needs-review, unverified, and non-public listings.
- - Pay metrics with fewer than 10 qualifying roles are shown as insufficient rather than a number.
- - Listings can change after they were last checked.
Machine-readable data: /api/ai-evaluator-stats.json