BigLaw Bench leaderboard
BigLaw Bench
3 models tested · Updated 2026-03-05 · Verified sources only
GPT-5.4 leads at 91.0%
1
GPT-5.4
OpenAI · Blog/OpenAI · 2026-03-05
Best legal document analysis. Structures complex transactions, maintains accuracy across lengthy contracts.
91.0%
2
Claude Opus 4.7
Anthropic · Blog/Anthropic · 2026-04-16
High effort eval. Better reasoning calibration on review tables and smarter handling of ambiguous document editing tasks.
90.9%
3
Claude Opus 4.6
Anthropic · Blog/Anthropic · 2026-02-05
Legal document analysis benchmark by Harvey. Strong professional reasoning capability.
90.2%
benchmark.space canonical public ledger of AI capability