University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

claude-sonnet-4-5-20250929

Anthropic
average
54.0
leaderboard rankof 114
100
benchmarksof 12
12
per 1,000 documentsmean over its benchmarks
$22
seconds per inputmean response time
22.4
Compare with other models Its row in the Matrix All its test runs

By task

  • Data correction

    63.4

    Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.

  • Document understanding

    28.6

    Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.

  • Information extraction

    66.5

    Best of 107 models with all its benchmarks: grok-4.6, 90.8.

  • Transcription

    45.8

    Best of 114 models with all its benchmarks: gemini-4-argon, 96.2.

By benchmark

BenchmarkScoreThe best current run on each benchmark, on the benchmark's metric scaled to 0–100. ± is the 95% sampling margin, from resampling the run's inputs.RankRank among all current models on that benchmark; ≈ for models that cannot be told apart within sampling error.Per 1,000 docss / inputBest run
Bibliographic Data 97.0±1.7 ≈29of 115 $38$0.19 run 24.96 T0225 · 24 Jan 2026prompt.txt
Blacklist Cardsolder run 86.3 ≈87of 114 $9.73$0.32 run 5.88 T0325 · 24 Oct 2025prompt.txt
Book Advert XML Repair 43.2±8.6 ≈16of 119 $15$0.77 run 10.88 T0468 · 8 Dec 2025prompt.txt
Business Letters 49.0±7.9 ≈100of 114 $13$0.75 run 3.34 T0226 · 24 Jan 2026prompt.txt
Company Listsolder run 66.9 96of 114 $25$0.37 run 20.93 T0379 · 28 Oct 2025prompt.txt
Duty Rosters 0.0±0.0 ≈105of 107 $36$0.76 run 93.71 T1279 · 24 Jul 2026prompt.txt
Fraktur Advertsolder run 0.0 ≈108of 114 $44$0.22 run 60.32 T0229 · 1 Oct 2025prompt_optimized.txt
General Meeting Minutes 81.3±11.6 ≈25of 108 $24$0.22 run 17.44 T1216 · 3 Jul 2026prompt.txt
Library Cards 85.0±1.8 ≈51of 114 $11$2.96 run 4.68 T0230 · 24 Jan 2026prompt.txt
Magazine Pages 2.2±4.3 ≈79of 115 $8.70$0.40 run 3.04 T0787 · 23 Mar 2026prompt.txt
Medieval Manuscriptsolder run 53.9 ≈93of 114 $11$0.14 run 13.43 T0293 · 24 Oct 2025prompt.txt
Personnel Cards 83.5±18.6 ≈78of 114 $27$1.65 run 10.29 T0581 · 16 Sep 2026prompt.txt

Price

FromInput, $ per million tokensOutput, $ per million tokensSource
30 Sep 20253.015.0archived price page

Every run

  1. T1279Duty Rosters17 Sep 20260.0$0.76
  2. T1216General Meeting Minutes17 Sep 202674.3$0.22
  3. T0787Magazine Pages16 Sep 20262.2$0.40
  4. T0581Personnel Cards16 Sep 202683.5$1.65
  5. T1279Duty Rosters24 Jul 20260.0$0.76
  6. T1216General Meeting Minutes3 Jul 202681.3$0.22
  7. T0787Magazine Pages23 Mar 20262.2$0.40
  8. T0581Personnel Cards9 Feb 202680.6$1.65
  9. T0468Book Advert XML Repair24 Jan 202640.3$0.78
  10. T0380Company Lists24 Jan 202659.4$0.36
  11. T0379Company Lists24 Jan 202662.1$0.39
  12. T0325Blacklist Cards24 Jan 202670.9$0.37
  13. T0293Medieval Manuscripts24 Jan 202628.5$0.16
  14. T0230Library Cards24 Jan 202685.0$2.96
  15. T0229Fraktur Adverts24 Jan 20260.0$0.090
  16. T0228Business Letters24 Jan 2026 · 32 signed letters53.0$0.46
  17. T0227Business Letters24 Jan 2026 · 25 unsigned letters48.0$0.29
  18. T0226Business Letters24 Jan 2026 · all 57 letters49.0$0.75
  19. T0225Bibliographic Data24 Jan 202697.0$0.19
  20. T0468Book Advert XML Repair8 Dec 202543.2$0.77
  21. T0380Company Lists28 Oct 202560.5$0.33
  22. T0379Company Lists28 Oct 202566.9$0.37
  23. T0325Blacklist Cards24 Oct 202586.3$0.32
  24. T0293Medieval Manuscripts24 Oct 202553.9$0.14
  25. T0229Fraktur Adverts1 Oct 20250.0$0.22
  26. T0228Business Letters1 Oct 2025 · 32 signed letters47.0$0.35
  27. T0227Business Letters1 Oct 2025 · 25 unsigned letters39.0$0.23
  28. T0226Business Letters1 Oct 2025 · all 57 letters44.0$0.58
  29. T0225Bibliographic Data1 Oct 202596.1$0.16
  30. T0230Library Cards30 Sep 202579.6$2.77