01 — CODING
General Software Engineering
Feature requests and bug fixes in real open-source codebases, each with hidden tests and a verified reference solution.
Data Catalog
Improve your models today with datasets for software engineering, general agentic workflows, and finance.
02 — AGENTIC
Long-Horizon Tool Use
Extended tasks across real tool ecosystems that require dozens of calls, state tracking, and recovery, with performance graded on outcomes.
03 — FINANCE
General Finance Knowledge
Research and financial modeling tasks grounded in real filings and market data.
04 — CUSTOM
Have a domain that isn't here? We build datasets and environments to spec. Reach out for more information.