micro1 is the largest gig-listing platform in the AI training space by volume — at any given time roughly 390 active opportunities sit on micro1.ai/experts/opportunities, spanning generalist evaluators, code reviewers, audio recording contributors, niche professional specialists (pixel artists, insurance analysts, nursing specialists), and engineering roles paying north of $100/hr. If you only apply to one platform in 2026, micro1 is the highest-yield first move for most people because the acceptance bar for generalist work is realistic and the pipeline of new listings is relentless.
Below is what the application actually looks like in 2026, where the pay clusters, and the math on whether the $3,000 referral program is worth promoting.
What micro1 hires for
Listings break into roughly four buckets, each with its own pay band:
- Generalist evaluators — reviewing AI-generated writing, ranking model outputs, light annotation. $20–$40/hr. The bulk of openings.
- Domain specialists — insurance, manufacturing, nursing, accounting, supply chain, etc. $20–$65/hr depending on credentials. Each "specialist" listing maps to a real client project with their own scoring criteria.
- Engineering and technical — software development evaluators, ML engineers, infrastructure specialists. $50–$150/hr. Smaller pool of listings, very fast pipeline once you're in.
- Audio and creative — pixel artists, voice actors, music annotators. Variable rates ($15–$70/hr typical) but very quick onboarding because the qualification is sample-based.
Pay is hourly with weekly invoicing. Most contracts are project-based (weeks to months) with rolling extensions if the client likes your work.
The application process
Application has three steps and the AI interview is where most candidates fail or pass:
- Profile + resume. Upload your CV; micro1 parses it and surfaces you against listings whose criteria match. Specificity beats breadth — listing five professional domains hurts your score for any of them. Pick one or two and own them.
- AI-led interview. A 15–25 minute structured video interview with an AI agent. The questions probe the experience on your resume in detail. Vague answers get a low score; concrete examples with metrics, customers, or technical specifics pass. Our getting-accepted playbook covers prep tactics that move the needle here.
- Listing-specific assessment. Once you clear the interview, individual listings may add a short skill check (a coding exercise, a writing sample, a domain quiz). Pass-through rates are high if you've made it this far.
The $3,000 referral program
micro1's referral program is structurally different from Mercor's earnings-share model and worth understanding:
- Flat $3,000 per successful referral who is hired and completes 10+ hours of paid work. Not a percentage; a fixed payout.
- No referral cap — you can refer as many people as you want.
- Anyone can participate. You don't have to work on micro1 yourself to refer someone, which is unusual.
- Payment lands in your micro1 wallet; withdraw via your preferred method. Sign up at refer.micro1.ai to get your code.
The math: a single high-pay-tier referral converts to $3,000 in your pocket. Three converts to a meaningful side income. Compared to Mercor's earnings-share (which ramps slower but keeps paying), micro1's flat payout favors front-loaded conversion.
Where micro1 fits vs. the rest
We break down the full landscape in our 2026 best-platforms guide, but the short version is:
- Mercor wins on top-end pay — domain-expert listings regularly hit $120–$200/hr. Smaller listing pool (~50 active).
- micro1 wins on listing volume and acceptance rate for non-specialists. Top end is lower (~$150/hr engineering) but there's always work.
- Handshake AI Fellowship wins for current grad students and new PhDs — up to $125/hr, U.S. work-authorization required.
The strategy most active contributors use: apply to all three in parallel after their first one accepts them. The interview practice compounds, and listing diversity smooths out the bursty nature of any single platform.
Mistakes that get applicants rejected
- Over-broad profile. A profile claiming expertise in software, finance, design, marketing, and law gets matched to generalist listings at $25/hr. The same person with a single tightly-positioned domain gets matched to specialist listings at $60–$150/hr.
- Abandoning the AI interview. If you start the interview and don't finish, the system marks the application incomplete. Block out 30 minutes when you're ready, not when you're between meetings.
- Ignoring the engagement type. Listings labeled "full-time" expect 40 hours/week; "part-time" is usually 15–25. Saying you'll do 40 hrs and then disappearing burns the listing.
Bottom line
For most people in 2026, micro1 is the best first stop. The volume of listings means there's almost certainly a match for your background; the AI interview is the only real gate; and the flat $3,000 referral payout is unusually generous compared to peer platforms. If you have specialized credentials, run micro1 in parallel with Mercor for the higher pay ceiling.
Browse all current micro1 listings filtered by pay range and commitment on our micro1 listings page, or compare across all platforms on the home page.