Definição
A fixed set of test questions used to compare models, with a published score for each. Experts are hired to write questions hard enough that current models get them wrong.
Used in: Expert roles in maths, science, law, medicine and coding.
Nas vagas abertas agora
411 of the open roles we list mention “benchmark” in the title or description, on 34 platforms (as of 3 November 2026). Of those that state an hourly rate (135 distinct postings), the median is $65/hr, and the middle half pay $50 to $90.
A maioria em: Turing (60), Mercor (51), Mindrift (43), Sepal AI (39), Alignerr (32), Invisible (23) e mais 28.
PhD Biology Expert / Life Sciences ExpertTuring·Local não informado·Período integral·Doutorado
Pagamento não informadohá 1 dia
US Patent Attorney — Prosecution & LitigationMercor·United States·Meio período
Os 10% mais bem pagos recebem:$180/hhá 3 dias
Licensed Physician (MD/DO) - Freelance AI TrainerMindrift·13 países·Horário flexível·Exige licença de advogado nos EUA (US bar)·MD
$100/hhá 4 dias
Database Systems Researcher (Research Coding Tasks)AfterQuery·Local não informado·Horário flexível·Doutorado
Pagamento não informadohá 5 dias
Management Consulting Expertmicro1·31 países·Meio período
$60–120/hhá 6 dias
Business Owners - Freelance AI Trainer ProjectJobgether·Saudi Arabia·Horário flexível
$100/hhá 7 dias
