Is Turing Legit? AI Training & Expert Jobs Review (2026)

Turing is a legitimate company, and its reviews are split down the middle

Turing is not a scam. The company raised a $111M Series E in March 2025 at a $2.2B valuation, and its listings describe work with frontier AI labs on training data, evaluations, and engineering. Contractors are engaged as independent contractors, paid in USD, and never asked to pay to apply.

Its public reputation is more mixed than that summary suggests. As of September 2026, Trustpilot shows a TrustScore of 2.5 across 201 reviews, with 67% five-star and 25% one-star ratings and very little in between. Contractors who get placed on a project tend to rate it highly for pay and flexibility. The low ratings come from people who passed screening and then waited for months, or who hit problems withdrawing their earnings, which the payments and pros-and-cons sections below cover.

Turing hires far beyond software engineers now

Turing's listings describe it as an AGI infrastructure company, which in practical terms means it supplies the people who produce and review training data for AI labs. Software engineering is still its largest category, but most of its open listings are for other fields. Healthcare roles ask doctors and nurses to check AI medical answers, legal and finance roles ask practitioners to write and grade domain-specific reasoning, and a steady stream of STEM, business, language, and voice roles fills out the rest.

Those non-coding roles are why Turing now shows up in searches from people who would never have considered it a few years ago. The engagement model is the same across all of them: you pass a screening, get matched to a client project, and work a set number of hours a week for the length of the contract.

Non-coders face an analytics and writing test, not a coding challenge

If you apply for a finance, healthcare, or language role, you will not be asked to write Python. Non-coding applicants are screened with an automated analytical challenge and a writing assessment, both of which Turing names in its listings. The analytical part tests whether you can reason through a problem in your field, and the writing part tests whether you can justify an answer in clear English, since that justification is most of what a client project will pay you for.

Engineers and data scientists still get the traditional route of timed coding problems in their stack, with system-design questions for senior roles. For both tracks, the assessments come before any contact with a person, and some listings add a short video interview after them. Passing does not guarantee work, because the next step is waiting for a client project to select your profile.

Turing does not publish hourly rates

Turing's listings state no hourly rate. Of the 359 Turing listings we track, 1 states a rate in the description text. The hourly figures on our Turing job cards are our own estimates, built from the role type and the country and labeled "estimate" on every card. Our pay methodology explains how they are made. Treat them as a planning range, not a quote.

The rate for a contract appears in the offer after a client selects you. Reviewers outside the US repeatedly describe Turing's USD rates as above local salaries, and healthcare, legal and senior engineering roles ask for more experience than language, writing and voice roles, which is the pattern our estimates follow.

To see how these rates compare with the rest of the market, the finance, medical, and biology AI training guides cover every platform hiring in each field.

Payments arrive in USD, though recent reviews report withdrawal holds

Turing engages you as an independent contractor, not an employee, and pays in USD. That arrangement is standard across AI training platforms, and it means no tax is withheld for you; the contractor tax guide covers what to set aside if this is your first contract role.

Recent one-star reviews describe withdrawals held for weeks or months with unresolved errors, along with referral bonuses that did not match what the dashboard showed. Five-star reviewers who are placed on active projects describe payouts arriving without trouble. The practical takeaway is to withdraw regularly once earnings clear rather than letting a large balance build up, and to keep your own record of hours and rates for each contract.

The PST overlap and contract length are where people get tripped up

Most Turing contracts ask for a 4-hour daily overlap with Pacific time, usually on top of a 20-hour weekly minimum. Of the 125 listings where we could extract a timezone rule, 85 require a partial PST overlap. The overlap exists so you can respond while the US research team is online. From India or Kenya that lands in your late evening, and from Europe it runs into the night, so price that into the hourly rate before you apply.

The roles are also contractor engagements with no health insurance or paid leave, and they end when the client project ends. Contract length is the variable most people overlook, and it depends heavily on the domain.

How Turing compares to Mercor, SME Careers, Alignerr, and Micro1

Turing competes for the same expert talent as Mercor, SME Careers, Alignerr, and Micro1. Turing publishes no rates, so its pay cell is blank; its advantage is the volume of non-coding listings in countries the others serve less. The pay column for the other platforms is the median top of band across each platform's live listings, estimates and placeholder ranges excluded.

For a ranked breakdown across every platform we track, see the best AI training platforms compared.

What reviewers praise, and where Turing falls short

The same pattern runs through both columns: contractors on an active project are happy, and contractors waiting for one are not. This summary draws on Turing's Trustpilot reviews as of September 2026.

Apply if you can cover US hours and wait a few weeks for a match

Turing is a good fit if you have engineering, STEM, healthcare, legal, or finance experience, can work four hours inside the Pacific day, and have other income while you wait to be matched. Treat passing the assessments as joining a queue, keep applying elsewhere in the meantime, and withdraw your earnings regularly once you are on a project.

It is a poor fit if you need paid work to start within days, or if a nightly overlap with California would not work for you. How to become an AI trainer covers platforms with faster entry.

Frequently asked questions about Turing