Blog

Benchmarks, Evals,
and Datasets

Research on AI capabilities in real-world workflows. We build rigorous evaluations and training environments for frontier models.

Introducing trajectories.sh and computer-1

A model-agnostic computer-use harness inside Harbor, and a trajectory viewer supporting multimodal computer-use agent trajectories

An RLVR dataset curated from Gauntlet, delivered in Harbor format for reinforcement learning.