University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

grok-4.20-0309-reasoning

xAI
average±2.1 (95%)
83.4
leaderboard rankof 114
≈14
benchmarksof 12
12
per 1,000 documentsmean over its benchmarks
$19
seconds per inputmean response time
53.8
Compare with other models Its row in the Matrix All its test runs

By task

  • Data correction

    79.2

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

  • Document understanding

    85.6

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

  • Information extraction

    85.1

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

  • Transcription

    87.8

    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 97.6±1.9 ≈8of 115 $35$0.18 run 104.77 T0718 · 26 Sep 2026prompt.txt
Blacklist Cards 94.1±2.7 ≈14of 114 $11$0.35 run 29.97 T0725 · 16 Sep 2026prompt.txt
Book Advert XML Repair 62.5±10.1 ≈1of 119 $29$1.46 run 45.57 T0728 · 23 Mar 2026prompt.txt
Business Letters 63.0±6.4 ≈24of 114 $12$0.70 run 22.94 T0719 · 16 Sep 2026prompt.txt
Company Lists 86.5±10.3 ≈29of 114 $16$0.25 run 49.32 T0726 · 16 Sep 2026prompt.txt
Duty Rosters 81.7±3.4 ≈30of 107 $27$0.58 run 95.21 T1313 · 16 Sep 2026prompt.txt
Fraktur Adverts 96.8±1.1 ≈27of 114 $31$0.15 run 101.59 T0722 · 16 Sep 2026prompt_optimized.txt
General Meeting Minutes 83.6±7.8 ≈25of 108 $19$0.17 run 66.17 T1274 · 16 Sep 2026prompt.txt
Library Cards 89.3±1.3 ≈15of 114 $9.13$2.40 run 27.41 T0723 · 16 Sep 2026prompt.txt
Magazine Pages 79.2±9.4 ≈18of 115 $9.80$0.45 run 27.1 T0813 · 16 Sep 2026prompt.txt
Medieval Manuscripts 70.8±14.3 ≈27of 114 $15$0.18 run 45.81 T0724 · 16 Sep 2026prompt.txt
Personnel Cards 95.9±2.2 ≈30of 114 $16$0.98 run 29.79 T0729 · 16 Sep 2026prompt.txt

Price

FromInput, $ per million tokensOutput, $ per million tokensSource
22 Jul 20261.252.5archived price page
22 Mar 20262.06.0archived price page

Every run

  1. T0718Bibliographic Data26 Sep 202697.6$0.18
  2. T1313Duty Rosters16 Sep 202681.7$0.58
  3. T1274General Meeting Minutes16 Sep 202683.6$0.17
  4. T0813Magazine Pages16 Sep 202679.2$0.45
  5. T0729Personnel Cards16 Sep 202695.9$0.98
  6. T0728Book Advert XML Repair16 Sep 202649.7$0.72
  7. T0727Company Lists16 Sep 202673.4$0.22
  8. T0726Company Lists16 Sep 202686.5$0.25
  9. T0725Blacklist Cards16 Sep 202694.1$0.35
  10. T0724Medieval Manuscripts16 Sep 202670.8$0.18
  11. T0723Library Cards16 Sep 202689.3$2.40
  12. T0722Fraktur Adverts16 Sep 202696.8$0.15
  13. T0721Business Letters16 Sep 2026 · 32 signed letters63.0$0.40
  14. T0720Business Letters16 Sep 2026 · 25 unsigned letters63.0$0.27
  15. T0719Business Letters16 Sep 2026 · all 57 letters63.0$0.70
  16. T0718Bibliographic Data16 Sep 202697.4$0.15
  17. T1313Duty Rosters24 Jul 202680.3$0.59
  18. T1274General Meeting Minutes3 Jul 202682.0$0.095
  19. T0813Magazine Pages23 Mar 202649.1$0.69
  20. T0729Personnel Cards23 Mar 202682.4$1.94
  21. T0728Book Advert XML Repair23 Mar 202662.5$1.46
  22. T0727Company Lists23 Mar 202634.7$0.50
  23. T0726Company Lists23 Mar 202649.3$0.62
  24. T0725Blacklist Cards23 Mar 202693.5$0.57
  25. T0724Medieval Manuscripts23 Mar 202651.5$0.31
  26. T0723Library Cards23 Mar 202686.1$4.68
  27. T0722Fraktur Adverts23 Mar 20264.7$0.19
  28. T0721Business Letters23 Mar 2026 · 32 signed letters32.0$0.57
  29. T0720Business Letters23 Mar 2026 · 25 unsigned letters18.0$0.36
  30. T0719Business Letters23 Mar 2026 · all 57 letters22.0$0.94
  31. T0718Bibliographic Data23 Mar 202682.9$0.42