Order-to-cash · finance workload

apply.cash.v1

Apply 20 customer payments to an open AR ledger of 30 invoices: multi-invoice payments, valid and invalid early-pay discounts, short-pays, overpays, loosely formatted remittance memos, and payments referencing unknown invoices (15 scenarios).

No published results for this workload yet — results appear here with the next edition.

Rate-card unit

Apply 1,000 remittances — rates on the rate card are the cost per successful outcome scaled to this unit.

Pass criteria

At least 18 of the 20 payments fully correct — the exact invoice set and the correct application code. Grading is deterministic; a model output either passes or fails, and infrastructure errors are excluded and rerun rather than counted against the model.

Prompt (verbatim)

Every model receives exactly this prompt with the item's fields substituted for the {placeholders} — no model-specific tuning.

You are applying customer payments (cash application) to the open
accounts-receivable ledger below.

Terms: invoices marked "2/10 net 30" allow a 2% early-payment discount when
payment is made within 10 days of the invoice date (day 10 inclusive);
otherwise the full amount is due within 30 days. Round each invoice's
discounted amount to the cent (round half up) before summing.

For each payment, identify the open invoice(s) it pays from the memo and the
payer, then code the application:
  paid_in_full — the amount equals the sum of the referenced open invoices
  discount_valid — paid within the discount window and the amount equals the
    discounted sum
  discount_invalid — the amount equals the discounted sum but the payment is
    past the discount window
  short_pay — the amount is less than the amount due and is not a discounted
    sum
  overpay — the amount is more than the sum of the referenced invoices
  no_match — the memo does not identify any open invoice (use invoice_ids [])

Memos may format invoice numbers loosely (case, spacing, missing prefix); an
invoice is referenced if its number can still be identified unambiguously.

Respond with ONLY a JSON object mapping every payment id to its application,
e.g. {{"PMT-99-01": {{"invoice_ids": ["INV-99-0007"], "code": "paid_in_full"}}}}.
Include every payment exactly once.

Open invoices (CSV):
{invoices_csv}

Payments received (CSV):
{payments_csv}

Dataset

Datasets are generated synthetically from seeded generators; the seed is frozen per edition and held out until the run window closes. Generator source hash (SHA-256): 9742ca5c68521c6968c754b2d51e59252095d530df5f2f4492559166b4a2b409. See the methodology §7 for the verifiability protocol and any edition page for the raw graded logs.