Benchmark
Company Knowledge Bench
Evaluate retrieval over varied company documentation and knowledge.
Published leaderboard
Explore scores and evaluation settings on the publisher’s leaderboard.
Open leaderboardCoverage
1,000 production-query cases and seven retrieval implementations in the published comparison.
Methodology
Score completeness and precision against Boolean criteria for preferred source chunks. Labeling agents are calibrated on 170 human-labeled cases.