Christine Copeland
Email text (the model got exactly this text)
From: Christine Copeland <christine.copeland@white.example> To: support@garrison-young.example Date: Thu, Jan 15, 2026 at 8:23 AM Subject: Contract termination Hi, We hereby terminate our service agreement and ask you to confirm the termination. The order number is 10530969. The 193 USD we paid should be refunded. This is completely unacceptable and very frustrating. Please resolve this by January 17. Thanks, Christine Copeland White Inc
Verdict
Cheapest without an error: GPT-6 Luna at $0.13 per 1,000 documents.
- Claude Sonnet 5 is also error-free but 29.4× more expensive.
- Mistral Small 4 is 1.7× cheaper, but: wrong value (Priority).
- Gemma 4 31B is 1.2× cheaper, but: wrong value (Priority).
Prompt given to every model
You triage the support inbox. Read the email below and return only JSON with these keys:
category one of: billing (invoices, payments, charges), delivery (where is the order, shipping,
address change), complaint (damaged, wrong or defective goods, warranty claims),
technical (login, app or system errors, integrations), cancellation (cancel an order,
contract or subscription), sales (quotes, prices, interest before buying)
priority one of: urgent, high, normal, low (rules below)
sentiment one of: positive, neutral, negative (the customer's tone)
customer_name full name of the customer who wrote the request
order_number the customer's order number, if mentioned
invoice_number invoice number, if mentioned
amount the amount of money the request is about (number)
currency ISO 4217 code of that amount (USD, EUR, CZK, …)
deadline the date by which the customer wants a resolution or answer (YYYY-MM-DD)
Priority rules:
- urgent: the customer cannot work right now, or the deadline is at most 1 day after the email date
- high: the deadline is at most 7 days after the email date, or the customer threatens to cancel
or escalate (lawyer, ending the business)
- low: the customer says explicitly that it is not urgent
- normal: everything else
Rules:
- A thread: judge the newest message; order numbers, invoice numbers and amounts may come from the
quoted earlier messages. A forwarded email: the customer is the original sender, not the
colleague who forwarded it.
- Resolve relative deadlines ("by Friday", "by tomorrow", "by the end of the month") against the
date of the message that states them.
- Support ticket numbers, tax IDs and phone numbers are not order or invoice numbers.
- If a value is not in the email, return null. Do not guess.
- Amounts as numbers with a decimal point, no thousands separators and no currency symbol.
How we score
Exact match per field after normalising dates, amounts and whitespace. Missing diacritics count as an error. Cost is the real token spend of this run, converted at the rate shown in the footer.
What each model extracted
| Field | Correct value | Mistral Small 48 / 9 fields | Gemma 4 31B8 / 9 fields | GPT-6 Luna9 / 9 fields | DeepSeek V4.1 Flash9 / 9 fields | Qwen3.8 Flash9 / 9 fields | Gemini 3.5 Flash Lite8 / 9 fields | Mistral Medium 3.58 / 9 fields | Gemini 3.8 Flash9 / 9 fields | GPT-6 Sol9 / 9 fields | Claude Sonnet 59 / 9 fields |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Category | cancellation | cancellation (correct) | cancellation (correct) | cancellation (correct) | cancellation (correct) | cancellation (correct) | cancellation (correct) | cancellation (correct) | cancellation (correct) | cancellation (correct) | cancellation (correct) |
| Priority | high | urgentwrong value | urgentwrong value | high (correct) | high (correct) | high (correct) | urgentwrong value | urgentwrong value | high (correct) | high (correct) | high (correct) |
| Sentiment | negative | negative (correct) | negative (correct) | negative (correct) | negative (correct) | negative (correct) | negative (correct) | negative (correct) | negative (correct) | negative (correct) | negative (correct) |
| Customer name | Christine Copeland | Christine Copeland (correct) | Christine Copeland (correct) | Christine Copeland (correct) | Christine Copeland (correct) | Christine Copeland (correct) | Christine Copeland (correct) | Christine Copeland (correct) | Christine Copeland (correct) | Christine Copeland (correct) | Christine Copeland (correct) |
| Order number | 10530969 | 10530969 (correct) | 10530969 (correct) | 10530969 (correct) | 10530969 (correct) | 10530969 (correct) | 10530969 (correct) | 10530969 (correct) | 10530969 (correct) | 10530969 (correct) | 10530969 (correct) |
| Invoice number | — | — (correct) | — (correct) | — (correct) | — (correct) | — (correct) | — (correct) | — (correct) | — (correct) | — (correct) | — (correct) |
| Amount | 193.00 | 193 (correct) | 193.00 (correct) | 193 (correct) | 193 (correct) | 193 (correct) | 193.00 (correct) | 193 (correct) | 193 (correct) | 193.00 (correct) | 193.00 (correct) |
| Currency | USD | USD (correct) | USD (correct) | USD (correct) | USD (correct) | USD (correct) | USD (correct) | USD (correct) | USD (correct) | USD (correct) | USD (correct) |
| Deadline | 2026-01-17 | 2026-01-17 (correct) | 2026-01-17 (correct) | 2026-01-17 (correct) | 2026-01-17 (correct) | 2026-01-17 (correct) | 2026-01-17 (correct) | 2026-01-17 (correct) | 2026-01-17 (correct) | 2026-01-17 (correct) | 2026-01-17 (correct) |
| Tokens | input / output | 559 / 83 | 585 / 82 | 683 / 128 | 824 / 234 | 909 / 450 | 573 / 102 | 559 / 83 | 573 / 330 | 683 / 96 | 819 / 225 |
| of which hidden reasoning | billed as output | 0 | 0 | 68 | 178 | 347 | 0 | 0 | 267 | 34 | 106 |
| of which from cache | input billed at a discount | 406 | 453 | 0 | 586 | 768 | 0 | 0 | 0 | 0 | 0 |
| Cost of this document | this run | $0.0001 | $0.0001 | $0.0001 | $0.0002 | $0.0002 | $0.0004 | $0.0015 | $0.0017 | $0.0023 | $0.0039 |
| Per 1,000 documents | for the same kind of document | $0.08 | $0.11 | $0.13 | $0.21 | $0.24 | $0.43 | $1.46 | $1.67 | $2.33 | $3.89 |
| Latency | this run | 1.2 s | 2.9 s | 2.7 s | 2.9 s | 13.9 s | 0.7 s | 0.9 s | 1.6 s | 2.5 s | 4.4 s |