ModelVerdict

Margaret Mills

English (US)Text7 fields

Document en-p0033 · layout hr · 7 fields · public part of the dataset · Same invoice in Czech → · Report an error in this example

Text (the model got exactly this text)

Employee record – Thomas-Andrews

Employee: Margaret Mills
Date of birth: July 25, 1989
SSN: 922-41-6247
Home address: 0172 Darren Springs Apt. 771, Staffordberg, FM 94707
Phone: 521-555-0008
Personal email: margaret.mills@inbox.example

Hired as staff accountant, 90-day probation period.
Reports to: Rachel Hall.

Verdict

Cheapest without an error: Gemma 4 31B at $0.10 per 1,000 documents.

  • Claude Sonnet 5 is also error-free but 35.5× more expensive.

Prompt given to every model

Find all personal data of natural persons in the text below (for anonymisation) and return only
JSON with these keys, each a list of strings (an empty list when there is none):

person_names       full names of people, each person once, as written in the text
email_addresses    email addresses of people
phone_numbers      phone numbers of people, as written
postal_addresses   home or delivery addresses of people, as written (street, city, ZIP)
birth_dates        dates of birth (YYYY-MM-DD)
national_ids       national identification numbers (e.g. SSN, birth numbers)
bank_accounts      bank account numbers of people, as written

Rules:
- Only data of natural persons. Company names, company registration or tax numbers (EIN, IČO,
  VAT), generic company addresses such as info@ or support@, order and ticket numbers are not
  personal data: leave them out.
- Dates that are not dates of birth (meetings, deliveries, start dates) are not personal data.
- Do not invent or complete values; copy them from the text.

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

FieldCorrect valueGemma 4 31B7 / 7 fieldsMistral Small 46 / 7 fieldsGPT-6 Luna6 / 7 fieldsQwen3.8 Flash7 / 7 fieldsGemini 3.5 Flash Lite7 / 7 fieldsDeepSeek V4.1 Flash6 / 7 fieldsMistral Medium 3.56 / 7 fieldsGemini 3.8 Flash7 / 7 fieldsGPT-6 Sol7 / 7 fieldsClaude Sonnet 57 / 7 fields
NamesMargaret Mills; Rachel HallMargaret Mills; Rachel Hall (correct) Margaret Mills; Rachel Hall; Thomas-Andrewsinvented value Thomas-Andrews; Margaret Mills; Rachel Hallinvented valueMargaret Mills; Rachel Hall (correct)Margaret Mills; Rachel Hall (correct) Thomas-Andrews; Margaret Mills; Rachel Hallinvented value Margaret Mills; Rachel Hall; Thomas-Andrewsinvented valueMargaret Mills; Rachel Hall (correct)Margaret Mills; Rachel Hall (correct)Margaret Mills; Rachel Hall (correct)
Email addressesmargaret.mills@inbox.examplemargaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)margaret.mills@inbox.example (correct)
Phone numbers521-555-0008521-555-0008 (correct)521-555-0008 (correct)521-555-0008 (correct)521-555-0008 (correct)521-555-0008 (correct)521-555-0008 (correct)521-555-0008 (correct)521-555-0008 (correct)521-555-0008 (correct)521-555-0008 (correct)
Addresses0172 Darren Springs Apt. 771, Staffordberg, FM 947070172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)0172 Darren Springs Apt. 771, Staffordberg, FM 94707 (correct)
Dates of birth1989-07-251989-07-25 (correct)1989-07-25 (correct)1989-07-25 (correct)1989-07-25 (correct)July 25, 1989 (correct)1989-07-25 (correct)1989-07-25 (correct)1989-07-25 (correct)1989-07-25 (correct)1989-07-25 (correct)
National IDs922-41-6247922-41-6247 (correct)922-41-6247 (correct)922-41-6247 (correct)922-41-6247 (correct)922-41-6247 (correct)922-41-6247 (correct)922-41-6247 (correct)922-41-6247 (correct)922-41-6247 (correct)922-41-6247 (correct)
Bank accounts—— (correct)— (correct)— (correct)— (correct)— (correct)— (correct)— (correct)— (correct)— (correct)— (correct)
Tokensinput / output377 / 121355 / 127460 / 177675 / 680365 / 174591 / 1,677355 / 127365 / 482460 / 2221,067 / 146
of which hidden reasoningbilled as output007951001,57303761280
of which from cacheinput billed at a discount24920305120366139000
Cost of this documentthis run$0.0001$0.0001$0.0001$0.0004$0.0005$0.0012$0.0015$0.0021$0.0031$0.0036
Per 1,000 documentsfor the same kind of document$0.10$0.10$0.13$0.35$0.54$1.16$1.49$2.08$3.14$3.59
Latencythis run4.5 s1.3 s2.1 s10.9 s1.3 s24.7 s0.9 s2.2 s4 s2.5 s