Vals AI
MortgageTax
Extracts information from mortgage tax certificates supplied as images: semantic fields (tax year, county, parcel number) and a numerical field (annualised amount due). The dataset is provided by Vontive.
About this benchmark
- Size
- 1,258 documents (20 public validation, 300 private validation, 938 test); reported scores use the test set
- Public split
- No public download; validation sets are available on request or for purchase, the test set is never shared
- Grading
- Accuracy. Models answer in JSON; regex is used when the format is not followed.
- Licence
- proprietary
- Submission
- Not stated
- Products allowed
- Not stated
- What it does not cover
- Archived as saturated: no longer run on new models. Document extraction only: no reasoning over filings, tools or market data.
Results
5 modelsEach score shows who reported it: the benchmark's authors (quoted from their paper), the benchmark's official board, or Moneda (run with the benchmark's own grader and rules). We do not copy other aggregators.
MortgageTax
accuracy %, higher is better. 5 models.
Updated 1 Sept 2026
- Claude Opus 572.1
- Claude Fable 5.170.8
- Claude Opus 4.770.3
- Claude Sonnet 570.0
- Claude Opus 4.869.9
MortgageTax
accuracy %, higher is better. 5 models.
Updated 1 Sept 2026
- 1Claude Opus 572.1
- 2Claude Fable 5.170.8
- 3Claude Opus 4.770.3
- 4Claude Sonnet 570.0
- 5Claude Opus 4.869.9
| 1 | Claude Opus 5 | Anthropic | 72.1 | test set (938 documents), archived board | Vals.ai MortgageTax leaderboard Official board, reported 1 Sept 2026 · checked 2 Nov 2026 |
|---|---|---|---|---|---|
| 2 | Claude Fable 5.1 | Anthropic | 70.8 | test set (938 documents), archived board | Vals.ai MortgageTax leaderboard Official board, reported 1 Sept 2026 · checked 2 Nov 2026 |
| 3 | Claude Opus 4.7 | Anthropic | 70.3 | test set (938 documents), archived board | Vals.ai MortgageTax leaderboard Official board, reported 1 Sept 2026 · checked 2 Nov 2026 |
| 4 | Claude Sonnet 5 | Anthropic | 70.0 | test set (938 documents), archived board | Vals.ai MortgageTax leaderboard Official board, reported 1 Sept 2026 · checked 2 Nov 2026 |
| 5 | Claude Opus 4.8 | Anthropic | 69.9 | test set (938 documents), archived board | Vals.ai MortgageTax leaderboard Official board, reported 1 Sept 2026 · checked 2 Nov 2026 |
Papers and links
Last checked 2 Nov 2026. Something out of date or wrong? Send a correction.