Frontier model A

Trial · allocation-flow-optimizer

← allocation-flow-optimizer
Outcome
FAIL1h 30m
Reward
0.0%
Checks passed
32 / 36
Runtime
1h 30m
Harness
Cursor CLI
Task version
v3.1.4

Checks

CheckResultCreditWeight
baseline: the optimizer answers a solvePASS100%1%
baseline: the app records an allocation runPASS100%1%
rank: timeline first takes on-time supply over cheaper late supplyPASS100%9%
rank: timeline first, then the cheapest on-time supplyFAIL0%9%
rank: each order gets its own optimumPASS100%6%
rank: cost first is strict, not a weightingPASS100%5%
rank: timeline first is strict, not a weightingPASS100%5%
rank: timeline first is optimised across requirementsPASS100%6%
rank: the ranking holds alongside substitutions, locks and supplier pinsPASS100%6%
vendor: the vendor-mix objective is gonePASS100%8%
summary: the solve reports the objectives in the order it optimised themPASS100%5%
summary: the solve reports each objective's solver statusPASS100%4%
persist: a reopened run returns its own solve summaryPASS100%11%
page: the planner ranks and turns off objectives, and the run uses that rankingPASS100%6%
page: a finished run shows the order it solved inPASS100%5%
scale: thousands of lots solve in secondsFAIL0%11%
preserve: POST /optimize: equipment types and substitutionsPASS100%0%
preserve: POST /optimize: supplier pinsPASS100%0%
preserve: POST /optimize: locked lotsPASS100%0%
preserve: POST /optimize: ready-date capsPASS100%0%
preserve: POST /optimize: cost capsPASS100%0%
preserve: POST /optimize: whole units within lot and requirement capacitiesPASS100%0%
preserve: POST /optimize: fractional supply lots left out and reportedPASS100%0%
preserve: POST /optimize: unfilled requirements and their reasonsFAIL0%0%
preserve: POST /optimize: demand filled before any objectivePASS100%0%
preserve: POST /optimize: single-objective plansPASS100%0%
preserve: POST /optimize: two ranked objectives, the lower breaking tiesFAIL0%0%
preserve: POST /optimize: response totals and assignment fieldsPASS100%0%
preserve: POST /optimize: refused requestsPASS100%0%
preserve: allocation runs (POST /api/allocate/runs)PASS100%0%
preserve: refused allocation runs (POST /api/allocate/runs)PASS100%0%
preserve: the run history and a reopened run (GET /api/allocate/runs)PASS100%0%
preserve: committing a run (POST /api/allocate/runs/[runId]/commit)PASS100%0%
preserve: a run's supply rows (POST /api/allocate/supply-rows with a runId)PASS100%0%
preserve: the supply selection resolved from a run (POST /api/allocate/runs, /api/allocate/phase-supply-selection)PASS100%0%
preserve: the Allocate page (GET /allocate)PASS100%0%