Long‑horizon RL environments for frontier AI labs.
Realistic, long-horizon environments for code, computer-use, and enterprise workflows that challenge SOTA models.
RL environments, custom & OTS.
Executable worlds with programmatic graders across three frontiers, hard enough that today’s best models fail most tasks.
Coding
Multi-file repos with build, run and test loops. Agents plan, edit, execute and repair across long sessions, graded against SWE-bench.
Computer-use
Full desktop and browser control, judged on the end state of long, multi-step tasks. Benchmarked on OSWorld.
Enterprise workflows
CRMs, spreadsheets, ticketing and finance — the real work companies run, with custom graders on your data.
What makes us different.
data quality Expert
network
Verifiable downstream model improvements
h-1, our 8B computer-use model, is trained entirely on our own computer-use environments — and ranks #9 on OSWorld, beside models many times its size.
300k+ expert network
Built on Huzzle.com. Our AI recruiter sources vetted specialists for any domain, on demand.
The result — hundreds of high-quality tasks per week. Thousands per month.
Request sample data.
Tell us what you’d like to see and we’ll tailor the sample to you.