# Harvey's Tenet model improves legal task success after post‑training

> Post‑training Harvey's Tenet on Fireworks AI lifts all‑pass rate on the Legal Agent Benchmark to 19.7%, about double the tasks completed.

Oossa · 2026-10-23 · https://oossa.com/en/harvey-s-tenet-model-improves-legal-task-success-after-post-training

Harvey teamed with Fireworks AI to post‑train its Tenet model on a Kimi K3 base using asynchronous reinforcement learning. The updated model finishes almost twice as many Legal Agent Benchmark (LAB) tasks and raises the all‑pass success rate to 19.7%, compared with 10.8% for the original Kimi K3. Tenet also scores higher on related agentic benchmarks without losing legal knowledge and runs at roughly the same per‑task cost ($5.92 vs $5.62).

## The facts

- Tenet all‑pass on LAB: 19.7% vs 10.8% for base Kimi K3
- Per‑task cost: $5.92 for Tenet, $5.62 for base model

## Why it matters

Law firms can get more complete legal drafts per dollar, reducing the time lawyers spend on repetitive document work.

## Sources & references

1. [Post-training Kimi K3 with Harvey for long-horizon legal work](https://fireworks.ai/customers/harvey-tenet-kimi-k3-post-training/) – Fireworks AI

Last updated: 2026-10-23