Data Extraction Race — results

Office Lease Agreement — Riverside Tower

$0.0053 total

Scoreboard

#1 Claude Opus 4.8Anthropic★ Won this round on this document
100/100
2285 ms · $0.0051
FieldExpectedModel answer
suiteSuite 1200Suite 1200exact
tenantPixel Forge StudioPixel Forge Studioexact
landlordRiverside HoldingsRiverside Holdingsexact
area_sqft45004500exact
lease_start2026-03-012026-03-01exact
term_months6060exact
monthly_rent97509750exact
permitted_usegeneral office and software developmentgeneral office and software developmentexact
late_penalty_pct55exact
security_deposit1950019500exact
#2 gpt-oss-20bOVH AI Endpoints (GRA)
97/100
10792 ms · $0.0001
FieldExpectedModel answer
suiteSuite 12001200close
tenantPixel Forge StudioPixel Forge Studioexact
landlordRiverside HoldingsRiverside Holdingsexact
area_sqft45004500exact
lease_start2026-03-012026-03-01exact
term_months6060exact
monthly_rent97509750exact
permitted_usegeneral office and software developmentgeneral office and software developmentexact
late_penalty_pct55exact
security_deposit1950019500exact
#3 Llama-3.1-8B-InstructOVH AI Endpoints (GRA)
90/100
2746 ms · $0.0000
FieldExpectedModel answer
suiteSuite 1200Suite 1200exact
tenantPixel Forge StudioPixel Forge Studioexact
landlordRiverside HoldingsRiverside Holdingsexact
area_sqft45004500exact
lease_start2026-03-01March 2026wrong
term_months6060exact
monthly_rent97509750exact
permitted_usegeneral office and software developmentgeneral office and software developmentexact
late_penalty_pct55exact
security_deposit1950019500exact

Source document

This Commercial Lease Agreement is made effective the first of March 2026 between Riverside Holdings (the "Landlord") and Pixel Forge Studio (the "Tenant"). The Tenant shall lease Suite 1200 at Riverside Tower, comprising 4,500 square feet of office space. Monthly rent is set at €9,750 due on the first business day of each month. A security deposit of €19,500 is required upon signing. The lease term is 60 months. Late payments incur a penalty of 5 percent. The permitted use is general office and software development. Extract the following fields as JSON: landlord, tenant, lease_start, suite, area_sqft, monthly_rent, security_deposit, term_months, late_penalty_pct, permitted_use.

How scoring works

Every model received the identical prompt and field list and answered once at temperature 0. We compare each field to the ground truth: an exact match earns full credit, the same value in a different format earns partial credit, anything else earns none. No model judges another — the score is pure field matching. A win means this model scored highest on THIS document with THESE fields, not that it is better overall.