Source · Vals AI Version · vals-ai snapshot 2026-10-03 Scores · 87
Test details
Visible tradeoffsThis is an objective signal, so it is mainly about measurable task performance rather than public taste.
source
Vals AI
metric
Accuracy (%)
judge
Objective
direction
higher better
group id
vals_mortgage_tax_current
domain
Document understanding
What it measures vs what it misses
✓ Measures
Reading and understanding tax certificates as images.
✗ Misses
Adjacent skills outside the benchmark task mix, latency, and cost.
Why this countsIt matters when the job is reading PDFs, tables, forms, or mixed-layout documents rather than plain chat.Same-test ruleThis percentile only compares models inside the exact benchmark/version group shown here. It is not a universal score.What it missesIt does not stand in for end-to-end document workflow quality or search relevance.
Leaderboard · this benchmark version
#1 · Claude Opus 5 (Adaptive Reasoning, Max Effort)
Fallback benchmark identity is visible for context but excluded from default ranking.
Identity
benchmark proxy (0.58)
Parsed from Vals AI BenchmarkView overall scores. Vals slug: mortgage_tax; provider: Anthropic. Backfilled from Claude Sonnet 4 via approved benchmark identity mapping map-claude-sonnet-4-5-to-4.