On-time delivery by hub, and why the late ones were late
The request
The task request supplied to the adapter, alongside the workspace below. Runtime instructions, authorized tools and skills also affect the system. Containerized runs withhold host-side references; local mode is not an isolation boundary.
I need August's on-time delivery by hub and a breakdown of why the late ones were late. Save it as delivery.xlsx with live formulas. Dispatch's notes explain how the stop file reads.
Workspace
- deliveries_august_2026.csv
- dispatch_notes.txt
Expected deliverables
- delivery.xlsx
This task also has a task-specific grader (check.py).
Recorded acceptance checks
The table uses check names and outcomes from the frozen campaign ledger. Acceptance requires all required predicates to pass; a predicate may itself contain a tolerance or coverage threshold. Its scope is part of the measurement contract.
The frozen scorer's check set differs from this task's source declaration. These are different evaluator versions. The counts below report only checks actually recorded; a missing check is never counted as a pass.
| Recorded check | Type | Required | Failed / observed |
|---|---|---|---|
| delivery.xlsx exists | file_exists | yes | 0 / 6 |
| live formulas | xlsx_has_formulas | yes | 0 / 6 |
| no error cells | xlsx_no_errors | yes | 2 / 6 |
| Portland on time | xlsx_value_present | yes | 2 / 6 |
| Denver shipments delivered | xlsx_value_present | yes | 2 / 6 |
| weather delays | xlsx_value_present | yes | 2 / 6 |
| on-time rates per hub | custom | yes | 2 / 6 |
Current task declaration (8 checks)
From task.yaml. Original runner grading and frozen retrospective scoring are separate. Read the frozen source for the rules behind the reported results.
- delivery.xlsx exists —
file_exists, required - live formulas —
xlsx_has_formulas, required - no error cells —
xlsx_no_errors, required - Portland on time —
xlsx_value_present, required - Denver shipments delivered —
xlsx_value_present, required - weather delays —
xlsx_value_present, required - total shipments delivered —
xlsx_value_present, required - on-time rates per hub —
custom, required
Recorded results, desk release
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes | Mean model estimate (USD) |
|---|---|---|---|---|---|
| Proto / DeepSeek V4.1 Flash | 3 | 3 | 11.0 | 0.1224 | |
| Codex / GPT-5.6 sol | 1 | 1 | 7.4 | 0.7742 |
These are repeated executions on the released fixture, not independent new business cases. Cost is captured model usage priced under the recorded assumptions, excluding infrastructure and human review. Per-attempt check verdicts, raw and frozen verdicts, usage, receipt and artifact hashes for this task are in results/latest/attempts.jsonl under "task": "delivery-performance".
Recorded results, desk comparison, 28 September 2026
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes |
|---|---|---|---|---|
| Proto / GPT-6 sol | 3 | 1 | 5.0 | |
| Codex / GPT-6 sol | 3 | 0 | 4.5 | |
| Proto / DeepSeek V4.1 Flash | 3 | 1 | 4.4 |
A separate campaign under the same frozen scorer, not pooled with the release. Conditions and data.