A Python library for evaluating agent harnesses and workflows with the model held fixed. It runs repeated trials, applies task checks, measures success, reliability, cost and latency, and compares harness designs.
Prices and medians update for the tier you select. Caution: thin market — treat statistics as indicative.
Compare at
harness-evals · Pro
No Pro plan
—
Pro-provider median
—
insufficient comparable pricing
vs Pro providers
No comparison
harness-evals doesn't sell Pro
ⓘ Comparisons are same provider type (provider) and same buyer tier (Pro). Never across tiers.
Market density
20th pctile
25 competing vendors
Comparable listings · 30d
—
Price spectrum · Pro plans · 1 of 2 providers priced · log scale · $49 → $49/mo
Platforms◻ shaded = middle 50% · line = median (all types)harness-evals has no comparable Pro price — not plotted
Products in this market
Ranked by how closely each one matches harness-evals's job. Prices show each provider's Pro state; entry prices are labelled as such. Unpriced products still belong to the market.