Mercor and micro1 are the two platforms that consistently surface at the top of "highest-paying AI training jobs" searches in 2026. Both run AI-led interviews, both place contractors with frontier labs and large enterprises, and both pay well above the annotation-platform average. They are also genuinely different products underneath the surface — different job types, different scheduling models, different ways of getting placed on work.
This is the honest head-to-head: where each one wins, where they overlap, and which to apply to first if your time is limited.
TL;DR — the verdict
- Mercor pays higher per-hour for top-tier domain experts (PhDs, senior engineers, M.D.s, J.D.s). The high end is $200+/hr; the typical placement is $80–$150/hr.
- micro1 has more total listings and a wider range of roles — generalist evaluation, coding, writing, analyst-style work. Typical pay is $40–$90/hr; the high end tops out around $150/hr.
- Mercor is harder to get into. The AI-led interview is the longest of the major platforms (45–90 min) and they reject most applicants.
- micro1 has lower friction. Faster interview, broader acceptance, more opportunities once you're in.
- You can absolutely do both. They don't have exclusivity clauses; many top contributors run both simultaneously and let the platforms compete for their time.
Pay: where the dollars actually land
Mercor
Mercor's published pay ranges per role are typically $60/hr at the bottom and $200+/hr at the top. The mode for accepted contributors is around $90–$120/hr. The genuinely high-end roles ($175–$200+/hr) are domain-expert work for frontier labs — evaluating model output on medical reasoning, legal analysis, graduate-level math, senior software architecture, finance and accounting at audit-partner level. You don't fall into these. You earn them by having verifiable senior-level credentials in a niche the labs need right now.
Mercor's payment schedule is weekly via Stripe. There's no minimum payout; whatever you logged in the prior week clears the following week.
micro1
micro1's pay scales are more horizontal — most roles fall in $40–$90/hr, with a long tail up to $150/hr for senior engineering and specialized expert work. The 10K+ contributor ceiling is lower than Mercor's individual peak, but the floor is also higher (very few sub-$30/hr roles).
micro1 pays weekly via Wise or Stripe. They tend to have more roles available at any given moment, which means once you're accepted, your hours can scale faster — you're not waiting for a specific niche to open up.
Per-hour vs total earnings
Here's the subtlety that most platform comparisons miss: higher hourly rate doesn't always mean higher total earnings. A Mercor placement at $150/hr that gives you 10 hours/week earns less than a micro1 placement at $80/hr that gives you 30 hours/week. Sustained hours matter as much as the rate.
In practice, Mercor placements are often spiky — intensive 2–6 week engagements at high pay, then a gap. micro1 placements tend to be more continuous. Pick based on what fits your schedule and income preferences, not just the per-hour headline.
Job types: where they actually overlap and diverge
Mercor's core competency: high-end SME evaluation
Mercor's bread-and-butter is matching contributors to frontier AI labs (OpenAI, Anthropic, Google DeepMind, xAI, and enterprise customers) for evaluation and red-teaming work that requires specific expertise. Typical engagements:
- Equity research analyst with 5+ years at a top-tier shop evaluating model output on financial reasoning
- Senior software engineer at FAANG-equivalent reviewing model-generated code for production-quality concerns
- Practicing physician evaluating medical reasoning chains in a specific specialty
- M&A attorney evaluating contract analysis output
- Mathematics PhD evaluating proofs at the research-publication level
micro1's core competency: scaled high-quality contractor placement
micro1 places contractors across a much broader band — from generalist evaluation (any educated adult with strong writing skills) up through senior engineering and specialist expert roles. Typical engagements:
- Generalist RLHF preference labeling on conversational AI output
- Code generation evaluation across mainstream languages (Python, TypeScript, Go, Rust)
- Analyst-style data work — review, summarization, fact-checking
- Subject-matter expert tasks across business, science, and professional domains
- Long-running engagements where the same contributor sticks with a specific client over months
Application process: what to expect
Mercor
You upload your résumé, the AI parses it, and you get a scheduling link for a video AI-led interview within 1–3 days. The interview is conducted by a conversational AI agent — no human screener. Expect 45–90 minutes of probing questions specifically targeted at the claims in your résumé. The AI will ask follow-up "tell me about the specific decision you made when…" questions designed to verify you actually did the work you claim. It will dig into technical details, ask you to walk through projects, and challenge you on specific aspects of your domain expertise.
After the interview, expect 5–10 business days for a placement decision. Most applicants are rejected. Those accepted are placed within their niche; you don't get to pick engagements.
micro1
Faster onboarding. Résumé upload → AI interview is typically 15–30 minutes, more conversational and less depth-probing than Mercor's. The bar is genuinely lower; most reasonable applicants with verifiable credentials make it through. Once accepted, you see available roles and can apply to specific ones rather than being placed.
For the application playbook that covers both, see our getting-accepted guide — the AI-interview tactics work for both platforms.
Which one should you apply to first?
- You have a verifiable senior credential (PhD in a hard science, 5+ years at FAANG or McKinsey/Bain/BCG, M.D., J.D., audit-partner-level accounting) → Mercor first. Your résumé probably qualifies you for $120–$200/hr placements. The interview is harder but the per-hour return is the highest in the market.
- You have real expertise but less prestige (mid-level engineer, 3–5 years of professional experience, specialist field without an advanced degree) → both — but micro1 first. micro1 is more likely to land you placements quickly while Mercor screens.
- You're a strong generalist (educated, careful, can write well, fast) → micro1 only. Mercor's model assumes domain expertise; generalists rarely place well there.
- You're a college student or early-career → neither yet. Both platforms expect a body of work. Build it first via the platforms in our no-experience guide.
Can you run both at the same time?
Yes. Neither platform has an exclusivity clause. The honest practical reality is that the highest earners on both platforms usually run multiple platforms simultaneously and accept whichever placement pays best at any given moment. Mercor contributors regularly carry a micro1 placement as their "stable income" while waiting for the next high-paying Mercor engagement to start.
Operationally: track your hours separately, manage tax implications by platform (each sends its own 1099-NEC), and be careful about non-compete or non-disclosure obligations on specific engagements — those are per-engagement, not platform-wide.
The honest weakness of each
Mercor's downside
Placement waiting time. After acceptance, you may wait 1–6 weeks for your first engagement, depending on what labs are sourcing in your niche. There's no way to apply directly to specific roles — Mercor's matching engine decides. If you need income starting next week, this is a problem.
micro1's downside
Pay ceiling is lower for the very top tier. If you're an audit-partner accountant or a research-publishing M.D., micro1 will pay you well — but Mercor will pay you 30–50% more for equivalent work. The price for micro1's faster onboarding is a lower individual peak.
How to actually apply to both efficiently
Submit to both on the same evening. Use the same résumé. The interviews are different enough that you'll have to engage with each separately; expect to budget 90 minutes for Mercor and 30 minutes for micro1 once they're scheduled. The two platforms don't share data — each one evaluates independently.
For the line-by-line playbook on the AI-led interview that both platforms run, see our AI training interview guide. For pay-tier context across the whole market, see our AI training pay breakdown.