Findings
- Open roles went from 4,130 to 5,862 (+1,732). 3,001 new roles were listed and 1,086 closed.
- At the end of the month the median hourly rate was $60, from 1,133 distinct postings that state a rate; the top 10% paid $150/hr or more.
- Most new openings: Lilt (+373), Ethos (+313) and Turing (+175).
- Fewest open roles compared with the start of the month: Weekday (-210), Alignerr (-33) and Welo Data (-27).
- Best-paid field at the median: Law, $100/hr from 86 postings.
By platform
| Platform | Open roles, start | End | Median $/hr (n) |
|---|---|---|---|
| Invisible | 560 | 557 | 35 (51) |
| Alignerr | 492 | 459 | 60 (101) |
| Lilt | 0 | 373 | – |
| Ethos | 0 | 313 | 90 (34) |
| Mercor | 277 | 305 | 85 (137) |
| Welo Data | 327 | 300 | 20 (79) |
| micro1 | 237 | 291 | 60 (133) |
| SuperAnnotate | 254 | 254 | – |
| Handshake AI | 228 | 245 | – |
| Prolific | 160 | 206 | 25 (16) |
| Anyone AI | 194 | 204 | 55 (17) |
| Turing | 0 | 175 | 30 (6) |
| TSMG | 0 | 160 | – |
| JT AI Labs | 0 | 159 | 60 (53) |
| RWS TrainAI | 98 | 152 | 17 (32) |
| Mindrift | 61 | 134 | – |
| Pasiflora AI | 132 | 132 | – |
| DataAnnotation | 112 | 112 | – |
| Rise Data Labs | 73 | 111 | 35 (29) |
| 10x Team | 0 | 107 | 145 (98) |
| Weekday | 308 | 98 | 105 (65) |
| Sigma AI | 97 | 86 | – |
| Perle | 0 | 63 | – |
| Volga Partners | 54 | 62 | 16 (21) |
| Claru | 60 | 60 | 45 (23) |
| Sepal AI | 0 | 59 | 60 (46) |
| Appen | 60 | 55 | 35 (6) |
| Lionbridge AI Data Training | 0 | 49 | 18 (29) |
| Babel Audio | 43 | 47 | – |
| Lifted, an Upwork Company | 0 | 46 | – |
| Braintrust | 47 | 34 | – |
| RemoExperts | 38 | 33 | 90 (27) |
| Terac | 18 | 33 | 90 (15) |
| Aptura | 0 | 27 | – |
| HumanSignal | 0 | 26 | – |
| TransPerfect DataForce | 27 | 24 | 30 (6) |
| xAI | 19 | 22 | 60 (7) |
| Chemin | 0 | 22 | – |
| CNTXT AI | 0 | 22 | – |
| TELUS Digital AI Community | 0 | 21 | 25 (8) |
| iMerit Scholars | 29 | 19 | 25 (5) |
| Sovrano AI | 21 | 18 | – |
| OneForma | 20 | 15 | – |
By field
| Field | Open roles | Median $/hr (n) | Top 10% from |
|---|---|---|---|
| Law | 192 | 100 (86) | 175 |
| Medicine & health | 329 | 100 (75) | 260 |
| Finance & accounting | 326 | 90 (92) | 160 |
| Other expertise | 1,150 | 80 (224) | 155 |
| Maths & STEM | 418 | 80 (63) | 145 |
| Coding | 484 | 80 (108) | 150 |
| Red-teaming & safety | 76 | 70 (22) | 110 |
| Writing | 86 | 55 (29) | 120 |
| Agents & RL tasks | 112 | 50 (18) | 75 |
| Data annotation | 288 | 40 (53) | 80 |
| AI evaluation & rating | 321 | 30 (95) | 100 |
| Voice & audio | 861 | 30 (134) | 100 |
| Languages & translation | 1,002 | 20 (88) | 45 |
| Robot training | 135 | 15 (18) | 35 |
Method and reuse
Figures count AI training roles only; paid research studies are left out. Pay figures count each distinct posting's hourly rate once, at the midpoint of its range, and leave out "up to" ceilings. More in the methodology. This report doesn't change once the month is over.
Free to reuse with credit under CC BY 4.0. Suggested credit: aitrainers.jobs, AI training jobs in October 2026, https://aitrainers.jobs/reports/2026-10. Current figures are on the pay data page and in open data.