ModelVerdict

Robert Lewis

Inglés (EE. UU.)Correo9 campos

Documento en-p0034 · diseño signature · 9 campos · parte pública del conjunto de datos · La misma factura en otro idioma: Checo → · Informar de un error en este ejemplo

Texto del correo (el modelo recibió exactamente este texto)

From: Robert Lewis <robert.lewis@wells-hayes.example>
To: support@gilmore.example
Date: Thu, Jan 15, 2026 at 12:02 PM
Subject: System problem

Dear support team,

The data export in your app fails with an error and the file is empty.

Thanks,

Robert Lewis
Purchasing Manager
Wells-Hayes
3048 Rhonda Branch Apt. 551
Nguyenton, RI 55244
Phone: (396) 560-7635
EIN: 51-9635789

This email and any attachments are confidential and intended solely for the addressee.

Veredicto

El más barato sin errores: Gemma 4 31B, a 0,10 € por cada 1000 documentos.

  • Claude Sonnet 5 tampoco comete errores, pero es 33,6× más caro.

Prompt enviado a todos los modelos

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.

Cómo puntuamos

Coincidencia exacta por campo tras normalizar fechas, importes y espacios. La falta de diacríticos cuenta como error. El coste es el gasto real en tokens de esta ejecución, convertido al tipo de cambio indicado en el pie de página.

Lo que extrajo cada modelo

CampoValor correctoGemma 4 31B9 / 9 camposGPT-6 Luna9 / 9 camposMistral Small 46 / 9 camposQwen3.8 Flash9 / 9 camposDeepSeek V4.1 Flash8 / 9 camposGemini 3.5 Flash Lite7 / 9 camposMistral Medium 3.59 / 9 camposGPT-6 Sol9 / 9 camposGemini 3.8 Flash9 / 9 camposClaude Sonnet 59 / 9 campos
Categoríatechnicaltechnical (correcto)technical (correcto)technical (correcto)technical (correcto)technical (correcto)technical (correcto)technical (correcto)technical (correcto)technical (correcto)technical (correcto)
Prioridadnormalnormal (correcto)normal (correcto) highvalor incorrectonormal (correcto)normal (correcto) urgentvalor incorrectonormal (correcto)normal (correcto)normal (correcto)normal (correcto)
Sentimientoneutralneutral (correcto)neutral (correcto) negativevalor incorrectoneutral (correcto) negativevalor incorrecto negativevalor incorrectoneutral (correcto)neutral (correcto)neutral (correcto)neutral (correcto)
Nombre del clienteRobert LewisRobert Lewis (correcto)Robert Lewis (correcto)Robert Lewis (correcto)Robert Lewis (correcto)Robert Lewis (correcto)Robert Lewis (correcto)Robert Lewis (correcto)Robert Lewis (correcto)Robert Lewis (correcto)Robert Lewis (correcto)
Número de pedido—— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)
Número de factura—— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)
Importe—— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)
Moneda—— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)
Plazo—— (correcto)— (correcto) 2026-01-22valor inventado— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)— (correcto)
Tokensentrada / salida625 / 57713 / 102602 / 67952 / 371854 / 389613 / 76602 / 57713 / 68613 / 590856 / 199
de ellos, razonamiento ocultose factura como salida0510294344001751495
de ellos, desde cachéentrada con descuento4530076858600000
Coste de este documentoesta ejecución0,0001 €0,0001 €0,0001 €0,0002 €0,0003 €0,0003 €0,0012 €0,0019 €0,0025 €0,0034 €
Por cada 1000 documentospara el mismo tipo de documento0,10 €0,11 €0,12 €0,20 €0,29 €0,34 €1,22 €1,94 €2,46 €3,41 €
Latenciaesta ejecución4,4 s2,4 s1 s5,6 s6,8 s0,6 s0,8 s1,8 s2,8 s3,8 s