Outlier AI Pay by Project Type: What Each Category Pays (2026)
Outlier AI's pay isn't one number. Software engineering, specialist expert work, voice evaluation, and general inbound roles pay differently. Real listing data, broken down by project type.
Every Outlier review on the internet gives you the same range: something like $15 to $55 an hour. That number is technically true and mostly useless, because it blends software engineers, French voice actors, and generalist reviewers into one bucket.
We pulled the live listings straight from Outlier's site on August 10, 2026, and broke pay down by what the project is. Below is what each category paid, using the exact figures in those listings, not rounded up or averaged across the platform.
Outlier's pay depends on which project you land, not one blended platform-wide number
- Software engineering runs $27.50-$50/hr across five active listings we pulled, the highest and most consistent band on the platform.
- Specialist expertise pays the most, up to $35-$75/hr for Law, but these roles are credential-gated and often country-restricted.
- Voice Conversation Evaluator pay swings roughly 4x by language: $31/hr for Korean down to $7.50-$8/hr for Spanish, for the identical task.
- General Inbound pays a flat $30/hr, but it's a holding queue, not a guaranteed project.
- Identical job titles can post at two different rates at the same time. Assessment score and degree level are the likely drivers, though Outlier doesn't state this directly.
Why "$15-$55/hr" Doesn't Tell You Much
Outlier posts dozens of projects at once, each with its own pay rate. A single blended range hides two things that matter more than the number itself:
Project type sets the floor and ceiling
Voice evaluation and software engineering are not competing for the same rate band. Knowing which category you're applying to tells you more than the platform-wide range ever will.
The same project can post at different rates simultaneously
We saw this directly in the data below: identical job titles, different pay, running at the same time.
Software Engineering
The highest and most consistent rates in the data we pulled. Five active listings, all coding-related, ranging from $27.50/hr to $50/hr.
| Project | Pay |
|---|---|
| Full Stack Software Engineer for AI Training | $27.50/hr |
| Full Stack Software Engineer for AI Training | $50/hr |
| Software Engineer for AI Training (Code Quality & Debugging Focus) | $27.50/hr |
| OpenClaw Agent Builder (Software Developers) | $50/hr |
| OpenClaw Agent Builder (Software Developers) | $35/hr |
Both the Full Stack Software Engineer listing and the OpenClaw Agent Builder listing appear twice, at different rates, with otherwise identical descriptions. Outlier doesn't publish why. The listings themselves note that PhD-level candidates and higher assessment scores can unlock higher rates, and that project need at the time you apply plays a role.
Specialist Expertise (Law, Math, Accessibility, Education)
This is where Outlier's top-of-range numbers come from. These are narrow, credential-gated projects, not open-application work.
Law Expertise
$35-$75/hr. Requires a J.D. or equivalent and practicing legal experience.
Math Expertise
$50/hr. Bachelor's or higher in mathematics, LaTeX proficiency expected.
Blind & Low Vision AI Evaluator
$35/hr. Restricted to people who are blind or have low vision and use assistive technology daily. US-based; Texas, Illinois, and Washington are excluded.
Qatar Educators
$50/hr. Requires current teaching experience in a Qatar Ministry of Education public school and fluent Arabic. Restricted to residents of Qatar.
Three of these four ($75/hr top-end law, $50/hr math and Qatar education) explain most of the high-end numbers you see quoted for Outlier. They also have the narrowest eligibility bars on the platform: real credentials, and in two cases, a specific country or lived-experience requirement.
Voice Conversation Evaluator (by Language)
Seven active language variants in the data we pulled, all under the same "Voice Conversation Evaluator" project. Same task, wildly different pay by language:
| Language | Pay |
|---|---|
| Korean | $31/hr |
| German | $29/hr |
| French | $25/hr |
| Italian | $20/hr |
| Simplified Chinese | $20/hr |
| Thai | $10/hr |
| Spanish | $7.50-$8/hr |
That's roughly a 4x spread for the same job description: live scenario-based conversations with an AI voice model, plus reviewing and scoring recorded conversations. Requirements are consistent across languages (native or highly fluent speaker, comfortable with unscripted conversation), so the pay gap tracks the language, not the difficulty of the work.
General Inbound / Open Interest
Two listings that aren't tied to a specific project. "General Inbound (Coding)" and "General Inbound" both pay $30/hr and function as an interest form: you apply without a matching project yet listed, and Outlier reaches out when something fits your background. One is coding-flavored, the other spans writing and reasoning work more broadly.
This is a reasonable fallback if nothing else on the list matches you, but it's a queue, not a guaranteed project.
What Moves Your Rate
Based on what Outlier states directly in its own listings (not speculation):
- Assessment score. Several listings note that skills assessment results affect the rate you're offered.
- Advanced degrees. Listings repeatedly call out that PhDs may be offered higher rates within the same project.
- Missions (surge pricing). Software engineering listings mention Outlier Contributors earning an average of 7.5% on top of core rates through Missions, with the top quartile of contributors seeing about 11% more.
- Project phase. Onboarding and overtime phases are explicitly called out as paying lower than core project work.
What we can't tell you from listing data is exactly why two identical-title listings post at different rates at the same time (Full Stack Engineer at $27.50 vs. $50, OpenClaw Builder at $35 vs. $50). It's consistent with assessment-based tiering, but Outlier doesn't spell that out publicly.
Frequently Asked Questions
Is $55/hr the normal rate on Outlier? ▼
No. $55/hr showed up once, as the top of a $35-$75/hr range for Law expertise specifically. Most listings we pulled sit between $20 and $35/hr. Treat any single headline number as the ceiling of one project, not the average across the platform.
Why does the same job title pay two different rates? ▼
We saw this directly: two 'Full Stack Software Engineer' listings, one at $27.50/hr and one at $50/hr, and two 'OpenClaw Agent Builder' listings at $35/hr and $50/hr. Same title, same description, different pay. Outlier appears to run the identical project at multiple rate tiers at once, likely tied to assessment results or project need at the time you apply.
Why is Spanish voice evaluation only $7.50-$8/hr when other languages pay $20-$31/hr? ▼
We can't say for certain from listing data alone, but the pattern in what we pulled shows a wide spread across voice languages: Korean and German sit near $30/hr, while Spanish and Thai sit under $10/hr. Supply of qualified speakers per language is the most likely driver, but that's inference, not something Outlier states.
Should I apply through General Inbound if I don't see my specialty listed? ▼
It's a real option and it pays $30/hr in the listings we saw, but it's explicitly a holding queue: you submit interest and Outlier matches you later if a project fits. If a specific listing matching your background exists, apply to that first.
Related guides
Outlier platform intel: full profile, requirements, and payout details.
Outlier empty queue decoded: what to do when the pay is good but the queue is empty.
Outlier alternatives: other platforms worth applying to alongside Outlier.
Best AI training platforms compared: how Outlier's pay stacks up against the rest of the market.
Cite this page
"Outlier AI Pay by Project Type: What Each Category Pays (2026)", aitrainer.work, aitrainer.work/guides/outlier-pay-by-project-type
MSc Human-Computer Interaction | Founder
Pietro is the founder and technical lead of aitrainer.work. He builds and maintains the platform's data pipeline, certification infrastructure, and editorial standards.