Agent: Custom Legal AgentsLLM: GPT-4, Claude 3.5#benchmark#legal-ai#agent-evaluation#llm-testing#legal-tech
Harvey LAB is a comprehensive benchmark for testing AI agents on realistic legal tasks. It includes 1,671 tasks across 24+ legal practice areas with execution harness, evaluation rubrics, and comparison dashboards for LLM agent performance assessment.