Mercor

Builds RL environments and datasets for AI labs.

Records how working professionals reason through real tasks and packages it as training data for frontier model developers.

Human evaluation, RLHF preference data and post-training datasets with a scalable-oversight focus.

Publicly listed veteran of the data industry offering training data, annotation and model evaluation with a large global crowd.

Crowd-voted public leaderboard for language, image and code models, with commercial evaluation services.

Data-curation research shop that builds environments and optimizes agents for enterprises and labs.