Request Samples

Sample task packets, benchmark holdout licensing, and environment evaluation access for AI researchers, data teams, and frontier labs.

Start with a sample, not a pitch

Most engagements start the same way: you tell us the capability you care about, we send the relevant cut. A sample packet includes representative tasks, their rubrics and verifiers, the expert baseline, and environment access so your team can run agents against it directly.

What we typically provide:

  • Sample task packets from any live occupation cohort
  • Evaluation access to running environments
  • Benchmark holdout licensing, shared or exclusive
  • Custom cohorts: name the occupation and workflow, we source the experts
  • VPC deployment for teams whose data cannot leave

We respond fastest to a short description of where your model is breaking and what a win would look like.

Research & data teams

Sample task packets, benchmark holdout licensing, environment evaluation access, and custom occupation cohorts.

Everything else

Press, partnerships, and enterprise deployments inside your own VPC.