Research

Research for faster model improvement

Benchmarks, papers, and field notes on where models fail, how to measure the gap, and what it takes to move the frontier.
Research areasUser SimulationVerifiers & RewardPost-Training & DistillationAdversarial RobustnessLong-Horizon AgentsCybersecurity

Benchmarks

Evaluations we build to find where models fail

Publications

Papers, preprints and collaborations with frontier researchers

Field Notes

Recent technical writing from our team

Orange twisted ribbon-like shape forming a loop on a black background.
Put the research to work

Turn a failure surface into
measurable model improvement.

Bring us the capability gap. We’ll shape the tasks, environments, and evaluation program needed to measure it—and improve it.

Talk to a Researcher