Legal AI platform Ivo has put seven frontier models through its new ContractBench benchmark, built on 350 real contracts. xAI's Grok 4.7 came out on top with a weighted pass rate of 58.3%, just ahead of Claude Opus 5.5 at 58.1% and Kimi K3 at 57.1% — a very tight race at the top.
Completion rates separated the leaders more clearly: Grok finished 99.5% of tasks and Kimi 100%, while Opus managed only 80.8%. Kimi was also 38% cheaper per completed rollout and 13% faster than Grok.
The most telling result is the ceiling itself: none of the seven models scored above 60%. AI can already speed up contract review considerably, but before anything is signed, human review remains indispensable.
Source: Read the original article
Comments