Show HN: FrontierHarness Eval – 9 harness, same model, cost per pass varies 17x
Summary
FrontierHarness Eval benchmarks 9 harnesses across 12 configurations on identical software tasks, highlighting significant variations in cost per task and pass rates. The article identifies leaders by quality, cost, and speed, and discusses methodological controls such as identical starts and neutral evaluation, along with caveats like cache effects and data exclusions.