LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models

LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models is an AI benchmark in the Benchmark Radar catalog.

Released: 2023-08-21
Openness: unknown
Importer: Benchmark Saturation

Source provenance

Original evidence

LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models results and reported scores

Reported benchmark scores · Benchmark Saturation

Reported scores: 5

Model	Organization	Reported value	Reported	Evidence
Claude Sonnet 4.5	—	83.2	—	evidence
Gemini 2.5 Flash Preview	—	82.8	—	evidence
Gemini 2.5 Pro Exp	—	83.6	—	evidence
GPT 5	—	84.5	—	evidence
Grok 4	—	83.4	—	evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.

Which sources cite LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models?