Mercor vs Handshake AI: Two Expert Tracks, Different Risks
Mercor vs Handshake AI compared on expert pay, eligibility, screening, and work stability. Mercor has the broader band; Handshake AI is US-only and less stable.
Pick
Mercor
Pick Mercor if you are outside the US, want the wider expert band, and can tolerate project churn.
Mercor pay guide
Pick
Handshake AI
Pick Handshake AI if you are a credentialed US expert who can pass an invite-only screen and wants the highest verified floor.
Handshake AI pay guideThe short answer
Mercor offers the broader expert band and does not publish a US-only requirement. Handshake AI holds the highest verified expert floor, but its US-only invite model and ban rate make it the riskier seat.
Best for
Mercor
Domain experts, engineers, and scientists; no published US-only requirement
Handshake AI
US-based credentialed experts in law, medicine, finance, and engineering
Published pay
Mercor
$50–$150/hr expert track; advertised average above $85/hr
Handshake AI
$60/hr floor and $100/hr ceiling for credentialed projects, averaging near $80/hr; generalist work about $17/hr
Eligibility
Mercor
No published US-only requirement; individual roles can restrict location
Handshake AI
US-only, invite-heavy, with verification through university records or transcripts
Screening
Mercor
AI video interview and trial project
Handshake AI
Invite, credential checks, and the R2I gating assessment introduced June 2026
Payment
Mercor
Weekly through the platform
Handshake AI
Weekly on a 7-day Stripe cycle; Deel on some long-term projects
Main risk
Mercor
Project cuts and the March 2026 breach
Handshake AI
Account bans, empty queues, and a mid-2026 project that paid 20–50% of earned hours
| Measure | Mercor | Handshake AI |
|---|---|---|
| Best for | Domain experts, engineers, and scientists; no published US-only requirement | US-based credentialed experts in law, medicine, finance, and engineering |
| Published pay | $50–$150/hr expert track; advertised average above $85/hr | $60/hr floor and $100/hr ceiling for credentialed projects, averaging near $80/hr; generalist work about $17/hr |
| Eligibility | No published US-only requirement; individual roles can restrict location | US-only, invite-heavy, with verification through university records or transcripts |
| Screening | AI video interview and trial project | Invite, credential checks, and the R2I gating assessment introduced June 2026 |
| Payment | Weekly through the platform | Weekly on a 7-day Stripe cycle; Deel on some long-term projects |
| Main risk | Project cuts and the March 2026 breach | Account bans, empty queues, and a mid-2026 project that paid 20–50% of earned hours |
Mercor and Handshake AI both sell access to expert AI work, and both pay above the generalist market. Mercor is the broader marketplace. Handshake AI is the narrower, US-only program with the highest verified floor and the least predictable work. The right choice starts with eligibility, not rate.
Pay: a wide band versus a high floor#
Mercor advertises $50–$150/hr for expert work, with an average above $85/hr. The band is wide because projects differ by client, credential, and task type. Handshake AI publishes a tighter expert band: a $60/hr floor and a $100/hr ceiling for credentialed projects, averaging near $80/hr. Generalist evaluation sits far lower at about $17/hr, and specialist evaluation near $40/hr.
| Track | Mercor | Handshake AI |
|---|---|---|
| Credentialed expert | $50–$150/hr advertised | $60–$100/hr, about $80/hr average |
| Technical or engineering | $50–$150/hr expert band | $75–$80/hr on public postings |
| Generalist evaluation | $16–$100+/hr project-dependent | About $17/hr |
| Payment schedule | Weekly | Weekly Stripe cycle |
Eligibility decides more than rate#
Mercor publishes no US-only requirement, though individual roles can restrict location. Handshake AI is US-only and invite-heavy. Verification runs through university records, transcripts, or a portfolio, followed by the R2I gating assessment introduced in June 2026. The assessment gives no score or feedback if you fail. If you are outside the US, the comparison ends here.
Screening and time to first work#
Mercor’s screen is an AI video interview and a trial project. Handshake AI starts with an invitation or application, then credential verification and the R2I assessment. Waitlists can run for months, and most projects cap at 10–20 hours per week once you are in.
Work stability is the real trade#
Mercor’s documented risk is project churn: a late-2025 contractor cut with lower-pay rehires, and a March 2026 supply-chain breach that exposed contractor records. Handshake AI’s risk is access. Its worker tracker recorded a 7.5% account ban rate and an 11.4% empty-queue mention rate across 1,654 discussions. A mid-2026 project paid workers 20–50% of earned hours. Tracked sentiment sits at 45/100.
Handshake AI pays premium rates with premium access risk. Never treat it as a sole income source, and keep your own hour log because activity tracking disputes are documented.
Who should pick which#
Pick Mercor if you are outside the US, want the wider rate band, and can absorb project cuts. Pick Handshake AI if you are a credentialed US expert who can pass an invite-only screen and wants the highest verified floor. If you are a US expert, applying to both is rational: Mercor for breadth, Handshake for the top band, with the expectation that neither guarantees next month.
See expert roles with published pay
NearSkill lists AI training and evaluation roles from these platforms with the pay and application route stated on each listing.
Written for real AI training and domain expert candidates. No fluff, no recycled job board advice.
Mercor vs Handshake AI, answered
Is Handshake AI better than Mercor?
Handshake AI has the higher verified floor for credentialed US experts: $60–$100/hr with an average near $80/hr. Mercor has the wider band, reaching $150/hr, and works with contractors outside the US. Handshake is the better seat only if you are US-based, credentialed, and can tolerate its ban and queue risk.
Can non-US workers join Handshake AI?
No. Handshake AI is US-only and invite-heavy, with identity verification through university records, transcripts, or a portfolio. Applicants outside the US should use Mercor, Micro1, or Turing instead.
How much does Handshake AI pay generalists?
Generalist evaluation sits around $17/hr, far below the credentialed expert band. Specialist evaluation is near $40/hr, and technical or engineering postings run $75–$80/hr. The $60–$100/hr band applies to credentialed expert projects, not entry-level work.
Is Handshake AI stable?
No. The platform is the least steady in this comparison. Its worker tracker recorded a 7.5% account ban rate, an 11.4% empty-queue mention rate, and offboardings without explanation. A mid-2026 project paid workers 20–50% of earned hours. Qualify on a second platform before relying on it.
Related comparisons
Mercor vs Micro1
Mercor pays the higher expert ceiling. Micro1 has the stronger payment record and a clearer engineering track. The right pick depends on your credential and how much screening you will accept.
Mercor vs Turing
Mercor pays more at the top and moves faster. Turing offers steadier volume, a lower median, and a daily US Pacific overlap. Here is the trade in numbers.
Handshake AI vs AfterQuery
Handshake AI holds the higher verified expert floor but restricts access to US applicants. AfterQuery pays $80–$200/hr for specialists, runs weekly Stripe payouts, and rejects tasks without explanation.
Stop comparing. Start matching.
Upload a resume and see live roles from these platforms ranked by fit, with the published pay attached. Free, no account.
Sources: Mercor: platform-published rates, CNBC (October 2025), and the 2026 platform review. Handshake AI: public job postings, the BreakingEven worker tracker (1 Jun–17 Aug 2026), Business Insider (April 2026), and the 2026 platform review. Rates are USD hourly and change with demand. Rates are indicative, not quotes. Payment timing and worker reports change; the linked platform pages carry the latest figures NearSkill can verify.