How much does DataAnnotation pay? Rates by track

DataAnnotation says its rates start at $20 an hour for translation work and run to $150 an hour or more for coding. For general chatbot-rating work it advertises $25 to $50 an hour or more. Those are starting rates the company publishes. It doesn't say what the average worker takes home.

That's what the company publishes. This post sets out those numbers, shows where they disagree with each other, and adds the little independent evidence there is on what people take home.

We don't list DataAnnotation jobs on our board, so our tracker has no hourly figures for it. We also earn nothing if you join. Every number below is from DataAnnotation's own pages, checked on 9 October 2026, or from a named source.

DataAnnotation pay by track

You pick a track when you sign up. DataAnnotation's FAQ gives a starting rate for each:

Track	Starting rate	What it covers (FAQ)
Multilingual	$20+ an hour	Translation, localisation
General	$25 to $50+ an hour	Evaluating chatbot responses
Coding	$40 to $150+ an hour	Code evaluation, debugging
STEM	$40 to $125+ an hour	Domain-specific work
Professional	$40 to $125+ an hour	Specialised work, such as law, finance and medicine

"Starting at" matters. Read each figure as the floor for its track, with a "+" on the end.

The numbers don't all agree

DataAnnotation puts several different figures on its own site:

• About page headline: "Starting at $45/hr".
• FAQ, General track: "Starting at $25-$50+ per hour".
• About page job cards: Professional Writer and Content Editor at "$25-$30+/hour". Localization Specialist at "$20-$50+/hour". Quantum Physicist, Investment Analyst, Applied Mathematician, Corporate Attorney, Senior Software Engineer and Clinical Researcher at "$50-$100+/hour".

So the $45 headline is not what a general-track newcomer should expect. Using the company's own cards, writing and editing work sits at $25 to $30 and up, and specialist work at $50 to $100 and up. The practical rule: read the rate on the task itself.

What your rate depends on

The FAQ says "Workers are added to projects based on expertise and performance." In plain terms, three things set what you earn:

1. Your track. Coding, STEM and professional work start at $40 an hour, against $20 to $25 for multilingual and general. They also need the skills and, for STEM, the degrees to match.
2. Your Starter Assessment. It's the way in for the track you pick, and you get one attempt. Our guide to passing the DataAnnotation Starter Assessment covers it.
3. Your performance. Projects are given to workers based on how well they've done.

The FAQ also says "Work is available 24/7/365" and "You choose which projects to accept based on your expertise". It does not promise any hours. Tasks come and go.

How and when you get paid

The FAQ says: "We send payments via PayPal. Deposits will be delivered within a few days after you request them." The About page says "Get paid after every project", with money you can withdraw whenever you want.

You need a PayPal account that works in your country. For a wider look at payout options, see how to get paid for AI training work outside the US.

You're an independent contractor, so no tax is taken out. Set money aside. How AI training pay works explains the basics.

What workers actually report

DataAnnotation doesn't publish what a typical worker earns, and I found no independent survey I'd trust. What exists is thin:

• TIME, 2 April 2024. One worker who had been there nearly two years said they were working on two projects, one at $20 an hour and one at $25, and earning about $400 to $500 a week. The same person said: "This is not permanent, as tasks come and go, but it is a great side income." That's one person, in 2024.
• Trustpilot, 9 October 2026. 3.9 out of 5 from 1,973 reviews. 54% give five stars, 19% four, 4% three, 2% two and 21% one star. Our 7 October read of the one-star reviews found them mostly about locked accounts, payouts on hold, never hearing back and no reply from support.

Both point the same way: the hourly rate is real for people who get work, but the weekly total depends on how many tasks you can get.

Where pay can go wrong

• Hours aren't guaranteed. A $50 rate and no tasks pays nothing.
• Unpaid time is disputed. In May 2025 a California worker sued Surge Labs, Inc. dba DataAnnotation (Cavalier v. Surge Labs, CGC-25-625502). The plaintiff's lawyers say workers do "off-the-clock work for training and project familiarization" and were misclassified as contractors. These are allegations. I found no ruling as of 9 October 2026.
• One try at the door. The FAQ says: "You can only take the Starter Assessment once. There are no retakes or second chances." If you fail, you don't get to see the higher rates.

Is it worth it?

If you have a skill that fits the coding, STEM or professional tracks, the published starting rates are $40 an hour or more. For general rating, the published start is $25 an hour, which is a fair side income as long as you accept that the work is patchy.

Don't wait on it. Apply to a second platform at the same time. Outlier AI vs DataAnnotation shows the two big names side by side, and the hiring tracker shows which platforms we list are hiring this week. For the full legit check, see the DataAnnotation review.

See DataAnnotation pay by track, and pay at every platform.

Sources checked on 9 October 2026

• DataAnnotation: FAQ, About and Trust & Safety
• Trustpilot: DataAnnotation
• TIME, Side Hustle or Scam? What to Know About Data Annotation Work (2 April 2024)
• Clarkson Law Firm, Clarkson represents Surge AI workers in labor law class action (20 May 2025)
• HiringRemote tracker, snapshot week of 5 October 2026, last checked 9 October 2026