The best-paid domains on Mercor in 2026

Rates on Mercor vary by domain more than by experience. A junior expert in a scarce domain often out-earns a senior generalist. Knowing which tracks pay the premium in 2026 is the most under-discussed lever for new evaluators. ## The current premium tier - Formal mathematics and theorem-proving (130-200 USD/hour) - Specialised programming languages with thin model coverage (OCaml, Rust systems work, Solidity formal verification): 90-150 USD/hour - Specific legal jurisdictions and regulatory work (US securities law, EU GDPR specifics): 100-180 USD/hour - Medical and clinical reasoning (validated credentials required): 100-180 USD/hour ## The mid tier - General LLM-agent trace evaluation (Python, web tooling): 60-100 USD/hour - Production ML systems reasoning (training, inference, distillation): 60-110 USD/hour - Data science and statistics: 50-90 USD/hour ## The entry tier - General code-quality pairwise evaluation in mainstream languages (Python, JavaScript): 35-60 USD/hour - Open-ended language comprehension and writing: 30-50 USD/hour ## What makes a domain pay well on Mercor Three traits compound: 1. Supply scarcity (few qualified evaluators globally) 2. Model weakness (the labs need help precisely where models still struggle) 3. Verifiable depth (the domain has clear right or wrong answers, not subjective taste) ## How to land a premium track Apply with the most specific declared expertise that is still truthful. "Senior backend engineer with 4 years on Stripe risk systems" qualifies for code-quality work. "Specialist in distributed consensus protocols who has read every Raft and Paxos paper" qualifies for niche systems-engineering tracks at the premium tier. ## Domains shifting in 2026 Mathematics and formal proof tracks have grown sharply in 2026 as labs push reasoning capabilities. Legal and medical tracks remain steady. General code-evaluation tracks have softened as more evaluators enter the pool. ## How to find out which tracks Mercor is actively recruiting Mercor publishes track openings on its careers page and on LinkedIn. The tracks listed with active recruitment are the ones with the highest current need; applying into those gets you screened fastest.