Historic feature — kept read-only for administrators. Nothing here can be changed.All historic features
Historic features
What the platform did before Bentham Studio 1.0. Kept as a read-only reference — all data is preserved, and nothing here can be changed.
Historic Studio
Probe tasks: undisclosed and disclosed runs, honesty labels, peer review.
Datasets
Benchmark questions authored by domain experts, and their peer review.
Benchmarks
Knowledge benchmarks run against frontier models, with graded answers.
Leaderboard
Model scores across benchmarks.
Commissions
Lab-funded commissions, time logs and payouts.
Categories
Benchmark categories.
EU AI Act compliance
Enterprise scoping assessments.