Mercor is being used by all of the top 5 AI labs and 6 of the Mag 7
Mercor provides the talent and data that measures and powers the most useful advances in AI.
Evaluate Your ModelPower-law experts generate power-law returns in training data
Elite
talent
marketplace
1M+
Experts
500k+
Interviews
75+
Expert NPS
Dual engagement model
As crowdsourcing models degraded in quality and transparency, Mercor emerged as the only vendor to combine elite talent, fully managed projects, full visibility into projects, and flexibility around tooling.
Market leader in rubrics
Mercor is on the cutting edge of evaluation and has been deeply focused on rubric development for 14+ months, partnering with leading frontier labs to refine processes for data collection, RL environments, and end-to-end rubric design.
Mercor Applied Research
The Mercor applied research team focuses on advancing frontier data with sophisticated evaluations and task designs. We work with partners to create novel, realistic evaluations and pioneer new data formats to power model breakthroughs.
APEX
Introducing APEX →
Bridging the gap between AI evaluation and economic value
RL
The Economy will become an RL environment →
The market for humans teaching models is based on the amount of tasks humans can do which agents can't do.
Evals
Welcome to the Era of Evals →
The primary barrier to applying agents to the entire economy is building evals for everything.
Data Types
The AI data layer for the frontier
Rubrics
Mercor is a pioneer in rubric development, having refined the process from sourcing to onboarding to data creation and review. We handle all modalities - text, image, audio, video, code.
RL Environments
We view RL environments in 3 parts: Scaffolding, creating mock applications; World Building, filling those applications with in-distribution data; and Task Creation, making high quality tasks within these simulated worlds.