F04
Models, efforts and routing
The source summary is shown as authored. Follow its evidence links for study-specific status and cohort limits.
Findings source section · FINDINGS.md
Hypothesis: H39–H50 ask whether model, effort, role allocation, or routing changes the completion-cost-time frontier. The evidence is panel-specific and does not establish a default model or route.
What ran: The family page links exploratory model and ladder records, historical model/effort rounds, and offline route diagnostics. R66 lacks its historical high-effort anchor cohort; R68b records an effort clamp; R72 and R74 remain separate.
Result: Numerator/denominator: not recorded for a complete matched model-and-effort frontier. Unit: official Build pass, cost, or wall time. Date: no common comparison date recorded. Status: interim, confounded, or withdrawn by round.
What it does not show: Offline route agreement and separate historical cohorts do not establish that escalation or task-text routing improves live Build outcomes.
Evidence: F04 hypothesis page, R66, R68b, R72, R74, Jev route context, Jev triage context, model screen, ladder inventory. Spec map: ladder recipe, Grok provider, and provider port.