# DataAnnotation vs Outlier vs micro1 vs Mercor vs Handshake AI vs Alignerr (2026) > DataAnnotation, Outlier, micro1, Mercor, Handshake AI, and Alignerr compared by pay, application effort, task supply, and best fit. - Canonical: https://paidtotrainai.com/ai-training-platforms-compared - Author: Luke Lashley (https://paidtotrainai.com/about) - Published: 2026-07-13 · Updated: 2026-08-06 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: https://paidtotrainai.com/disclosure I have an account on these platforms. They all pay you to improve AI models, but they target different people, pay different rates, and screen applicants differently. Here's the comparison, followed by the order I'd apply in. ## The short version | | DataAnnotation | Outlier | micro1 | Mercor | Handshake AI | Alignerr | |---|---|---|---|---|---|---| | **Typical pay** | $22.50–$30+ general; my coding work is $50–$75 | $15–$25 general, $22–$35 coding, to $100 expert | $20–$45 eval, $40–$95 expert, to $230 senior eng | $65–$90 expert, $100–$300 specialist | Project-specific hourly or per-task | Role-specific | | **Gate** | Reasoning or specialist assessment; my coding exam took about an hour | Auto-graded skill assessments | Skill-matched Zara interview; prior passes can shorten or skip it | AI interview per concept | Resume / academic profile | Role-specific application and assessments | | **Task volume** | Consistently high for my coding account; less reliable for some generalists | Most in the industry on paper — matching is the bottleneck | High | High, but responses can take months | Variable by project; vet terms before tasking | Best treated as targeted applications | | **Best for** | Programmers and people with a testable specialty | Only people who exactly fit an open targeted project | Domain experts and general data analysts | Mid-career/senior domain experts | Students, grad students, postdocs or general experts | Coding, robotics, finance, and Lua specialists | ## Who each platform is actually for **DataAnnotation** is my strongest recommendation for programmers. I normally have $50-per-hour coding work available, have personally worked at up to $75 per hour, and often find the tasks genuinely enjoyable. My entrance assessment took about an hour and I was accepted about a week later. The caveat is that my task supply may reflect my coding domain; some generalists report dry dashboards. Full breakdown: [DataAnnotation review](/dataannotation-review). **Outlier** is the biggest name and, on paper, has the most raw work in the industry — it's the contributor platform run by Scale AI. In practice it's the hardest place here to actually get a task: generalist rates run lower than the others, and the notorious "empty queue" hits even contributors with multiple verified skills. The only application worth making is into a specific project that matches you exactly — I keep a [targeted-project picker](/outlier-review#targeted) updated in my [Outlier review](/outlier-review) — and even then, treat whatever shows up as a bonus. **micro1** casts the widest net — trades, engineering, multilingual specialists, general annotation. You name your own expected rate, and the gate is a recorded Zara interview matched to the role's skills. Qualifying results from the last year can now shorten or skip later interviews, and listing volume is the highest in the space. Full breakdown: [micro1 review](/micro1-review). **Mercor** is for people whose resume already justifies reviewing hard AI answers: medical licenses, legal backgrounds, senior technical depth. Its reusable per-skill assessments probe real work history, and its top rates are the highest of the three. Full breakdown: [Mercor review](/mercor-review). **Handshake AI** is still a low-effort application: a quick form off your existing profile, no interview, and matching-project emails. It historically skews toward students and researchers, but current projects can be hourly or per task, with time caps and approval rules that can change the effective rate. Apply, then vet every offer before starting. Full breakdown: [Handshake AI review](/handshake-ai-review). **Alignerr** is a targeted-specialist play: coding, robotics, finance, accounting, and Lua roles each have their own application link. Do it when a role matches your actual background; skip it if you would be guessing. Alignerr has a lot of domain expertise roles, but I'd lean towards Mercor unless you have already exhausted Mercor. Full breakdown: [Alignerr review](/alignerr-review). ## The head-to-head matchups ### DataAnnotation vs Outlier For programmers, DataAnnotation wins decisively in my experience. I can generally find $50-per-hour coding work there, while Outlier's coding rates are commonly lower and its empty-queue problem makes availability unreliable even after passing assessments. DataAnnotation's initial coding exam took me longer, but acceptance produced real work. Outlier is still worth a targeted application when a specific project fits; I would not choose its generic pipeline over DataAnnotation. ### DataAnnotation vs micro1 DataAnnotation is the better open-dashboard fallback for me: once accepted and qualified, I can choose from coding tasks without applying to a new job each time. micro1 uses a role-specific application and may interview you on skills your recent results do not cover, but a landed micro1 engagement can run longer and its senior engineering ceiling is much higher. Programmers should use both: DataAnnotation for flexible task work and micro1 for longer contracts. ### micro1 vs Outlier On paper Outlier wins on raw volume; in practice micro1 wins on everything that matters. Outlier's auto-graded assessments are quick, but passing them doesn't mean tasks — even a [targeted application](/outlier-review#targeted) only improves your odds, and generalist pay runs $15–$25/hr when work does show up. micro1's Zara interview is more substantial, but recent qualifying results now carry forward: passed skills are skipped, and fully covered roles can be Fast-Tracked without another interview. The gigs you land also pay more ($40–$95/hr expert, up to $230 senior engineering), run longer, and are better managed. Prioritize micro1; add a targeted Outlier application only if a project is a dead-on fit. ### Outlier vs Mercor These sit at opposite ends. Outlier is the volume play in theory — the most raw work in the industry at the lowest typical rates — but the empty queue means that volume often never reaches you. Mercor is the credentials play: the highest ceiling of the four ($100–$300/hr specialist roles), but applications can sit for months. If your resume includes a license, an advanced degree, or senior technical depth, do Mercor's reusable assessments and let them work in the background — just don't count on Outlier to pay you while you wait. ### Mercor vs micro1 Mercor's ceiling is higher — premium specialist roles list $150–$300/hr against micro1's top engineering band of $100–$230/hr — but most expert roles on both pay $40–$150/hr, and micro1 responds to applications much faster. The gates differ more than the pay: Mercor's per-skill assessments are reusable across roles, while micro1's role-specific Zara interview now reuses passing skill coverage from qualifying interviews completed within the last year. Mercor suits mid-career and senior experts whose work history does the talking; micro1 casts a wider net and moves quicker. ## The strategy: stack them None of these platforms require exclusivity. The pattern that works — cheapest applications first. For resume tweaks, fresh listings, and application tracking, read [How to Apply for AI Training Jobs](/how-to-apply-ai-training-jobs). You can also browse the [current Micro1 and Mercor AI training jobs](/opportunities), sorted newest first. 1. **Start with Handshake AI** — no interview at all; the application runs off your resume and academic profile. Read the written pay basis, time cap, and approval rules before accepting a project. Apply to Handshake AI → 2. **If you code or have a strong specialist background, take DataAnnotation's assessment early.** My coding exam took about an hour and acceptance took about a week. Generalists can move this step later because task supply is less certain. [Read my DataAnnotation review and apply](/dataannotation-review). 3. **Do Mercor's assessments next** — they're per skill, not per role, so a few hours of assessments can be reused across many role applications. Start on Mercor → 4. **Add Outlier only if a targeted project matches you exactly** — the skill assessments are auto-graded in minutes, but the onboarding hours are unpaid and even a [matched application](/outlier-review#targeted) doesn't guarantee tasks. Skip the generic application entirely. 5. **Add Alignerr if one of the role-specific tracks fits**: coding, robotics, finance, accounting, or Lua. Pick an Alignerr role -> 6. **Then spend most of your applying time on micro1** — the first Zara interview can take focused time, but later applications can assess only missing skills or skip the interview when qualifying results from the last year cover the role. The gigs you land run longer, are better managed, and pay well. Apply to micro1 → ## The fine print Task availability fluctuates on every platform, and none guarantee full-time volume. These are freelance contracts: no benefits, and you handle your own taxes. This is supplemental income, not a job — and while demand is strong in 2026, nobody promises it stays that way. Get started: DataAnnotation · micro1 · Mercor · Handshake AI · Outlier (pick a project) · Alignerr (pick a role) ## Frequently asked questions **Which pays more, Mercor or micro1?** Mercor's ceiling is somewhat higher — premium specialist roles list $150–$300/hr versus micro1's top engineering band of $100–$230/hr — but on both platforms most expert roles pay $40–$150/hr, and micro1 responds to applications much faster. **Can I work on multiple AI training platforms at once?** Yes. None of these platforms require exclusivity, and many top earners stack them — steady weekly hours on one platform while waiting for higher-paying engagements on another. **Which AI training platform should a beginner apply to first?** Start with Handshake AI because it has no interview. Programmers and strong specialists should take DataAnnotation's assessment early; generalists can leave it until later because task supply is less predictable. Mercor's assessments are reusable, while micro1 can now shorten or skip Zara interviews when qualifying results from the last year cover a new role's skills. Add Alignerr or Outlier only when a targeted role matches you. **How does Outlier compare to Mercor and micro1?** Outlier, run by Scale AI, advertises the most raw task volume in the industry but pays less — $15–$25/hr generalist, $22–$35/hr coding, up to $100/hr expert — and has a chronic empty-queue problem: many contributors, even with verified skills, rarely see tasks. Mercor and micro1 pay experts more, and an accepted application there is far more likely to turn into actual paid work.