University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

claude-haiku-5-5

Anthropic
average±5.1 (95%)
67.5
leaderboard rankof 114
≈70
benchmarksof 12
12
per 1,000 documentsmean over its benchmarks
$0.87
seconds per inputmean response time
4.2
Compare with other models Its row in the Matrix All its test runs

By task

  • Data correction

    55.9

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

  • Document understanding

    55.5

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

  • Information extraction

    74.7

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

  • Transcription

    82.6

    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 Dataother prompt 76.7±37.6 ≈91of 115 $1.47$0.007 run 6.48 T1900 · 24 Oct 2026prompt_v2.txt
Blacklist Cards 88.9±3.6 ≈51of 114 $0.56$0.018 run 2.21 T1907 · 24 Oct 2026prompt.txt
Book Advert XML Repairother prompt 18.6±7.7 ≈64of 119 $0.74$0.037 run 5.14 T1909 · 24 Oct 2026prompt_v2.txt
Business Letters 57.0±6.6 ≈65of 114 $0.86$0.049 run 2.04 T1901 · 24 Oct 2026prompt.txt
Company Listsother prompt 78.1±14.1 ≈58of 114 $1.03$0.015 run 4.33 T1908 · 24 Oct 2026prompt_v2.txt
Duty Rosters 52.7±12.0 ≈87of 107 $0.52$0.011 run 3.89 T1913 · 24 Oct 2026prompt.txt
Fraktur Adverts 78.8±38.6 ≈32of 114 $1.38$0.007 run 9.09 T1904 · 24 Oct 2026prompt_optimized.txt
General Meeting Minutes 84.8±7.4 ≈25of 108 $1.06$0.009 run 4.68 T1912 · 24 Oct 2026prompt.txt
Library Cardsother prompt 84.7±1.5 ≈51of 114 $0.43$0.11 run 1.75 T1905 · 24 Oct 2026prompt_year-string.txt
Magazine Pagesother prompt 20.8±10.4 ≈56of 115 $0.52$0.024 run 2.09 T1911 · 24 Oct 2026prompt_explicit-json.txt
Medieval Manuscripts 75.8±14.8 ≈27of 114 $0.68$0.008 run 3.91 T1906 · 24 Oct 2026prompt.txt
Personnel Cards 93.1±2.3 ≈58of 114 $1.24$0.076 run 4.38 T1910 · 24 Oct 2026prompt.txt

Price

FromInput, $ per million tokensOutput, $ per million tokensSource
24 Oct 20260.10.5archived price page

Every run

  1. T1913Duty Rosters24 Oct 202652.7$0.011
  2. T1912General Meeting Minutes24 Oct 202684.8$0.009
  3. T1911Magazine Pages24 Oct 202620.8$0.024
  4. T1910Personnel Cards24 Oct 202693.1$0.076
  5. T1909Book Advert XML Repair24 Oct 202618.6$0.037
  6. T1908Company Lists24 Oct 202678.1$0.015
  7. T1907Blacklist Cards24 Oct 202688.9$0.018
  8. T1906Medieval Manuscripts24 Oct 202675.8$0.008
  9. T1905Library Cards24 Oct 202684.7$0.11
  10. T1904Fraktur Adverts24 Oct 202678.8$0.007
  11. T1903Business Letters24 Oct 2026 · 32 signed letters59.0$0.031
  12. T1902Business Letters24 Oct 2026 · 25 unsigned letters56.0$0.018
  13. T1901Business Letters24 Oct 2026 · all 57 letters57.0$0.049
  14. T1900Bibliographic Data24 Oct 202676.7$0.007