Kimi K3 Excels in Legal Benchmark Test

rohanpaul_ai · x · 2026-07-19

According to a tweet, Kimi K3 performed exceptionally well in a benchmark designed for autonomous legal work, scoring 26.7%—nearly double the 14.2% achieved by Claude Fable 5. The test includes 120 practical tasks across 24 legal domains (such as drafting memos and discovery summaries). Models must autonomously review case files and generate final legal documents under very strict evaluation criteria.

Related event: Kimi K3 Leads Harvey Legal Benchmark(4 posts)→

Original post →

More from Models

Models channel →