The Frontier of Legal Intelligence, Built in the Open.
Harvey is pushing the frontier of legal intelligence through open research. Our research agenda spans evaluation, benchmarking, post-training, harness optimization, verification, and agent behavior. We work alongside frontier labs, neolabs, AI startups, and academic researchers to bring agent research into Harvey. The long-term goal of Harvey is to contribute to the research and infrastructure required for the legal industry to create a frontier ecosystem.
Research Partners
Our research partners provide invaluable feedback and guidance on research direction, dataset quality, and collaborations to push the frontier of AI
Models
View allResearch on training and deploying models to make them maximally effective for legal work
Benchmarks
View allDatasets for training and evaluating models and agents. Leaderboards are hosted by our partners at Artificial Analysis and Vals.
Notes
View allResearch notes about models, agents, and applied AI
