Outlier Review (2026): Pay Rates, Empty Queues, and How to Actually Get Tasks
I actually work on these platforms — opinions come from my own dashboards and payouts. Some links are referral links that pay me at no cost to you. Full disclosure.
Outlier is the biggest name in AI training work — it’s the contributor platform run by Scale AI, the data company behind multiple major labs. However, it is pretty bad about leaving you in silence with no roles, even after you have verified multiple skills.
What the work is
Outlier contributors evaluate and improve AI model outputs: rating and rewriting model responses, solving domain problems (math, coding, science, medicine, law), and conversation tasks in dozens of languages. Client work spans major AI labs and enterprises.
Pay rates
| Work type | Typical range |
|---|---|
| General tasks | $15–$25/hr |
| Coding projects | $22–$35/hr |
| Specialized domains (math, science) | $25–$60+/hr |
| Expert review (legal, medical) | up to $100/hr |
Payouts run weekly on Tuesdays (for the previous Tuesday–Monday), via PayPal, AirTM, or ACH.
Know this going in: Outlier has been squeezing rates as supply grows. Several projects that paid $28–$35/hr in early 2025 were restructured to $18–$22/hr by 2026. Specialists have held their rates much better than generalists.
The application
You need to verify your identity, upload your resume, and then take assessments focused on skill areas. You should verify as many skills as possible.
How the skill assessments work
The flow I have experienced for the skill verification is:
- A short (1-minute) video to start, asking you a generic question about something like programming
- 1–2 questions about the domain with pretty clear-cut answers.
- 2–3 questions about the domain where it asks you to implement something properly and explain what was wrong.
You will be automatically graded in a matter of minutes.
The empty queue: Outlier’s defining problem
Search any Outlier forum and the top complaint is always the same: “EQ” — empty queue. You pass onboarding and then nothing shows up. Tasks only flow when you’re attached to an active project, and a generic application leaves the matching entirely up to Outlier’s pipeline.
The workaround is to not apply generically:
Apply through the link that matches you
This part matters more on Outlier than on any other platform: if you apply generically, you can sit in the queue with few or no tasks matched to you. Applying through a link targeted at a specific project puts you straight into that project's pipeline — if you fit the profile, you're far more likely to be onboarded and earning quickly.
-
US, Canadian, or UK cardiologists — attendings, fellows, or advanced residents — to review ECGs and solve complex clinical scenarios that train AI in diagnostic reasoning.
-
Analytical contributors with technical fluency, strong attention to detail, and experience reasoning through software, data, QA, or AI workflows, for OpenClaw-style AI agent training projects.
-
Valkyrie V10 (OB/GYN & Women's Health) →
US, Canadian, or UK MDs specializing in OB/GYN or Women's Health — attendings, fellows, or advanced residents — to solve clinical scenarios that train AI in diagnostic reasoning.
-
Native Korean, German, Bangla, or Chinese speakers for audio tasks. Must read/write English and have working computer audio; no pro hardware needed.
None of those you? Use the general application link — projects rotate constantly, and I update the targeted list above as new ones open.
These are referral links: Outlier pays me a bonus if you end up doing paid work. It costs you nothing and doesn't change your rate — details here.
The downsides
- Task volume swings hard. Even matched to a project, phases end and queues empty with no warning.
- Rate compression. The 2025→2026 restructurings cut some generalist projects by a third. Domain expertise is the only durable defense.
- Unpaid onboarding. The assessments take real hours before your first paid task.
Verdict: who should apply
Outlier makes sense for almost everyone if you enter through a project that fits you — it has the most work in the industry and the widest range of skill levels. Enter generically and you may wait weeks staring at an empty queue. If your background is strong enough for an interview-gated platform, run Mercor or micro1 in parallel — their rates run higher — and see the full head-to-head in Outlier vs micro1 vs Mercor vs Handshake AI.
Frequently asked questions
Is Outlier legit?
Yes. Outlier is operated by Scale AI, one of the largest AI data companies, and pays weekly via PayPal, AirTM, or ACH. It's the highest-volume platform in the space — the common complaint isn't payment, it's task availability.
How much does Outlier pay?
Roughly $15–$25/hr for general tasks, $22–$35/hr for coding work, $25–$60+/hr for specialized domains like math and science, and up to $100/hr for fields like legal and medical review. US generalists average around $30/hr.
Why am I not getting tasks on Outlier?
The 'empty queue' (EQ) is Outlier's defining problem: you only see tasks when you're matched to an active project. Generic applicants often wait weeks. Applying directly into a specific project that fits your background largely avoids this.
How and when does Outlier pay?
Weekly, on Tuesdays, for work completed the previous Tuesday through Monday (midnight UTC), via PayPal, AirTM, or ACH transfer.