Bill a hotel's group bookings across a mid-month rate rise, weekend nights, cancellations with time-zone offsets and a city-tax cap
Exact prompt
# Bill a hotel's November group bookings Inputs: `bookings.csv` (booking_id, guest, company, room_type, arrive, depart, guests, status, cancelled_at, checkout_time), `rates.csv` (room_type, from_date, weekday, weekend), `hotel.json` and `front-desk-note.txt`. Amounts are in pounds. `hotel.json`, `rates.csv` and these rules are the record; where the front-desk note disagrees, the rules win. Rules: - Each night of a stay (arrive up to the night before depart) is charged at the rate in force on that night's date (the row with the latest from_date on or before it); nights falling on a weekday listed in weekend_nights use the weekend rate. - A booking whose company is in corporates (ignoring letter case) gets corporate_discount_pct off its room charge or cancellation charge only, the discount rounded half-up to 0.01 per booking. - A stayed booking whose checkout_time is later than late_checkout_after pays late_fee; checking out exactly at late_checkout_after is free. - A cancelled booking is free if cancelled_at (which carries its own UTC offset) is at least free_cancel_hours before checkin_time on the arrival date in the hotel's timezone; otherwise, and for every no_show, the first night's room rate is charged. - City tax for a stayed booking = guests x nights x city_tax_per_person_night, counting at most city_tax_max_nights nights. - Write `out/billing.json`: `{"room": {"<booking_id>": <room charge after discount>, ...}, "late_fees": {...}, "cancellation_charges": {...}, "city_tax": {...}, "total": <sum of everything>}`. `room` and `city_tax` list every stayed booking; `late_fees` and `cancellation_charges` list only bookings charged one. ## How this run works - Work on your own: nobody will answer questions in this task, and writing to `ask/` counts against the run. - Write only under `out/`. Create `out/` if it is missing. Inputs are synthetic; don't fetch anything from the network. - Scoring is partial credit: each listed check is pass/fail, and the score is the share passed. Numbers are compared to within 0.005 unless the task says otherwise.
Bill a month of hotel bookings night by night. Rates rise mid-month and Friday and Saturday nights are dearer, corporate names appear in mixed letter case, checkout exactly at 11:00 is free, cancellations carry their own UTC offsets and land either side of the 48-hour line, and city tax stops after five nights. A front-desk note gets the rate, time-zone and late-checkout rules wrong. Original benchmark task written for usecaserank.com (not based on a specific public post). Headless, offline run on a reproducible fixture. Each run's notes say how its cost was measured (estimates are marked).
Success criteria: nightly_rates corporate_discount cancellations late_checkout city_tax_and_total
Best model for this task so far: Claude Haiku 5.5 (effort default (no --effort flag) · thinking on) (success, $0.0054, 23s).
All runs (6)
100% pass rate across 6 eligible runs: 6 pass, 0 partial, 0 fail. Partial and failed runs are listed below with everything else.ⓘ
| Model | Cost | Time | Step-insⓘ | Autonomy | Outcome | Trustⓘ | Date | Compare |
|---|---|---|---|---|---|---|---|---|
| GGemini 3.8 Flashmedium · thinking on | $0.21 | 1m | 0 | Fully autonomous | pass | Trust 80% | 2026-10-23 | |
| GGemini 4 Argonmedium · thinking on | $0.09 | 1m | 0 | Fully autonomous | pass | Trust 80% | 2026-10-23 | |
| OGPT-6.1 Soldefault · thinking off | $0.04 | 31s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-23 | |
| AClaude Opus 5.5default · thinking on | $0.13 | 23s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-23 | |
| AClaude Haiku 5.5default · thinking on | $0.0054 | 23s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-23 | |
| AClaude Sonnet 5.5default · thinking on | $0.06 | 14s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-23 |
Embed this result
A live badge for this use case. It links back here and updates as new runs are published.
Get new runs and ranking changes weekly
Which models got better or cheaper on real tasks like this one.