AppSci-Bench evaluates whether a model can elicit requirements, build an agent, create an internal evaluation, and generalize to held-out business tas
The Energy-Based Reasoners paper reports 345x less inference compute in a bounded comparison. The number is a measurement contract, not a universal sp
Agent-workdays are capacity signals. A practical measurement system separates throughput, validity, intervention, and durable research outcomes.
LifeVal tests personalized everyday tasks with artifacts and state diffs. Use a risk and verification ladder before an agent acts in the real world.