Harvey unveils Tenet, a model post-trained for long-horizon legal work

rhythmrg · x · 2026-08-21

Harvey shared the results of six months of research across benchmarking, post-training, and open-weight models, introducing Harvey Tenet, trained specifically for long-horizon legal work while generalizing across other agent and legal benchmarks.

New unlocked specializations include: M&A diligence (coordinating high-scale, long-horizon tasks in RLM harnesses), Review Tables (high-volume document review and structured data extraction), and Firm Knowledge (leveraging memory and emergent structured taxonomies of firm knowledge for better search).

Original post →

More from Models

Models channel →