University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

gpt-5.6-luna

OpenAI
average±1.9 (95%)
80.8
leaderboard rankof 114
≈25
benchmarksof 12
12
per 1,000 documentsmean over its benchmarks
$8.94
seconds per inputmean response time
14.9
Compare with other models Its row in the Matrix All its test runs

By task

  • Data correction

    67.8

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

  • Document understanding

    93.1

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

  • Information extraction

    84.2

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

  • Transcription

    80.8

    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 97.7±1.8 ≈8of 115 $20$0.10 run 16.38 T1375 · 25 Jul 2026prompt.txt
Blacklist Cards 87.4±3.3 ≈79of 114 $5.59$0.18 run 7.79 T1382 · 25 Jul 2026prompt.txt
Book Advert XML Repair 41.4±9.4 ≈16of 119 $1.55$0.077 run 16.37 T1385 · 16 Sep 2026prompt.txt
Business Letters 59.0±6.4 ≈24of 114 $1.63$0.093 run 4.62 T1376 · 16 Sep 2026prompt.txt
Company Lists 87.6±5.9 ≈29of 114 $1.99$0.030 run 9.77 T1383 · 16 Sep 2026prompt.txt
Duty Rosters 87.7±2.6 ≈18of 107 $29$0.60 run 31.09 T1389 · 25 Jul 2026prompt.txt
Fraktur Adverts 91.5±5.7 ≈32of 114 $4.45$0.022 run 31.22 T1379 · 16 Sep 2026prompt_optimized.txt
General Meeting Minutes 85.1±8.4 ≈25of 108 $14$0.13 run 14.66 T1388 · 25 Jul 2026prompt.txt
Library Cards 84.7±1.5 ≈51of 114 $4.53$1.19 run 4.6 T1380 · 25 Jul 2026prompt.txt
Magazine Pages 97.2±2.7 ≈1of 115 $6.21$0.29 run 9.3 T1387 · 25 Jul 2026prompt.txt
Medieval Manuscripts 56.5±14.0 ≈72of 114 $16$0.19 run 19.39 T1381 · 25 Jul 2026prompt.txt
Personnel Cards 94.3±2.9 ≈58of 114 $2.43$0.15 run 14.05 T1386 · 16 Sep 2026prompt.txt

Price

FromInput, $ per million tokensOutput, $ per million tokensSource
16 Sep 20260.21.2archived price page
25 Jul 202616archived price page

Every run

  1. T1375Bibliographic Data26 Sep 202697.0$0.019
  2. T1380Library Cards17 Sep 202683.0$0.23
  3. T1389Duty Rosters16 Sep 202686.5$0.11
  4. T1388General Meeting Minutes16 Sep 202683.2$0.029
  5. T1387Magazine Pages16 Sep 202697.2$0.059
  6. T1386Personnel Cards16 Sep 202694.3$0.15
  7. T1385Book Advert XML Repair16 Sep 202641.4$0.077
  8. T1384Company Lists16 Sep 202669.2$0.030
  9. T1383Company Lists16 Sep 202687.6$0.030
  10. T1382Blacklist Cards16 Sep 202687.3$0.037
  11. T1381Medieval Manuscripts16 Sep 202656.2$0.034
  12. T1379Fraktur Adverts16 Sep 202691.5$0.022
  13. T1378Business Letters16 Sep 2026 · 32 signed letters62.0$0.060
  14. T1377Business Letters16 Sep 2026 · 25 unsigned letters57.0$0.035
  15. T1376Business Letters16 Sep 2026 · all 57 letters59.0$0.093
  16. T1375Bibliographic Data16 Sep 202696.9$0.020
  17. T1389Duty Rosters25 Jul 202687.7$0.60
  18. T1388General Meeting Minutes25 Jul 202685.1$0.13
  19. T1387Magazine Pages25 Jul 202697.2$0.29
  20. T1386Personnel Cards25 Jul 202694.1$0.72
  21. T1385Book Advert XML Repair25 Jul 202640.2$0.41
  22. T1384Company Lists25 Jul 202676.9$0.15
  23. T1383Company Lists25 Jul 202686.5$0.16
  24. T1382Blacklist Cards25 Jul 202687.4$0.18
  25. T1381Medieval Manuscripts25 Jul 202656.5$0.19
  26. T1380Library Cards25 Jul 202684.7$1.19
  27. T1379Fraktur Adverts25 Jul 202690.3$0.11
  28. T1378Business Letters25 Jul 2026 · 32 signed letters59.0$0.28
  29. T1377Business Letters25 Jul 2026 · 25 unsigned letters61.0$0.16
  30. T1376Business Letters25 Jul 2026 · all 57 letters58.0$0.44
  31. T1375Bibliographic Data25 Jul 202697.7$0.10