Harvey releases Tenet legal model, boosts pass rates by 82%
dejavucoder · x · 2026-08-20
Harvey introduced Tenet, the first model post-trained specifically for legal work. Built on the Kimi K3 base and trained with FireworksAI, Tenet utilizes a mix of public legal data, synthetic data, and human expert simulations. It achieves an 82% increase in all-pass rates on the LAB benchmark relative to the base model, securing SOTA on LAB Contracts and second place overall. The gains also generalize to other agentic benchmarks, and the model is optimized for token efficiency.
Related event: Harvey Launches Harvey II Platform and First Self-Trained Legal Model Tenet(7 posts)→
More from Models
- Meta releases 816B Muse Spark Open, now the top US open-weights model at AA Index 39 — ArtificialAnlys · 2026-10-16
- Nolla Health Post-Trains Nemotron 3.5 VL: Visual Diagnosis Jumps from 35% to 66.8% — LoganGrasby · 2026-10-16
- Dev argues coding models now beat nearly all engineers in every branch of software — zack_overflow · 2026-10-16
- Prediction: Chinese frontier models could match GPT-6 by March 2027 — teortaxesTex · 2026-10-16
- Liquid AI's d1-ultra, #1 on Artificial Analysis Decision Index, lands on Vercel — JosephJacks_ · 2026-10-16
- Weekly AI roundup: Claude Haiku 5.5 cuts prices ~90%, Gemini 4 Argon output jumps to 1M tokens — DeepLearningAI · 2026-10-16