University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

gpt-6-sol

OpenAI
average±1.6 (95%)
86.3
leaderboard rankof 114
≈8
benchmarksof 12
12
per 1,000 documentsmean over its benchmarks
$19
seconds per inputmean response time
14.3
Compare with other models Its row in the Matrix All its test runs

By task

  • Data correction

    67.6

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

  • Document understanding

    95.9

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

  • Information extraction

    90.0

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

  • Transcription

    90.9

    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 96.6±1.6 ≈40of 115 $37$0.19 run 25.12 T1825 · 27 Sep 2026prompt.txt
Blacklist Cards 96.4±1.9 ≈14of 114 $9.66$0.32 run 5.1 T1832 · 27 Sep 2026prompt.txt
Book Advert XML Repair 37.5±11.3 ≈41of 119 $8.86$0.44 run 9.04 T1835 · 27 Sep 2026prompt.txt
Business Letters 68.0±7.3 ≈10of 114 $15$0.88 run 5.45 T1826 · 27 Sep 2026prompt.txt
Company Lists 97.7±1.9 ≈1of 114 $17$0.26 run 10.61 T1833 · 27 Sep 2026prompt.txt
Duty Rosters 94.8±1.5 ≈5of 107 $29$0.61 run 25.09 T1839 · 27 Sep 2026prompt.txt
Fraktur Adverts 97.6±0.3 ≈15of 114 $31$0.15 run 32.09 T1829 · 27 Sep 2026prompt_optimized.txt
General Meeting Minutes 86.0±7.9 ≈25of 108 $21$0.19 run 15.25 T1838 · 27 Sep 2026prompt.txt
Library Cardsother prompt 90.3±1.4 ≈1of 114 $7.21$1.89 run 6.36 T1830 · 27 Sep 2026prompt_year-string.txt
Magazine Pagesother prompt 95.3±4.4 ≈1of 115 $11$0.49 run 5.99 T1837 · 27 Sep 2026prompt_explicit-json.txt
Medieval Manuscripts 77.3±10.6 ≈27of 114 $22$0.26 run 20.72 T1831 · 27 Sep 2026prompt.txt
Personnel Cards 97.7±0.9 ≈1of 114 $20$1.22 run 10.2 T1836 · 27 Sep 2026prompt.txt

Price

FromInput, $ per million tokensOutput, $ per million tokensSource
27 Sep 20262.010.0archived price page

Every run

  1. T1839Duty Rosters27 Sep 202694.8$0.61
  2. T1838General Meeting Minutes27 Sep 202686.0$0.19
  3. T1837Magazine Pages27 Sep 202695.3$0.49
  4. T1836Personnel Cards27 Sep 202697.7$1.22
  5. T1835Book Advert XML Repair27 Sep 202637.5$0.44
  6. T1834Company Lists27 Sep 202675.3$0.24
  7. T1833Company Lists27 Sep 202697.7$0.26
  8. T1832Blacklist Cards27 Sep 202696.4$0.32
  9. T1831Medieval Manuscripts27 Sep 202677.3$0.26
  10. T1830Library Cards27 Sep 202690.3$1.89
  11. T1829Fraktur Adverts27 Sep 202697.6$0.15
  12. T1828Business Letters27 Sep 2026 · 32 signed letters65.0$0.56
  13. T1827Business Letters27 Sep 2026 · 25 unsigned letters85.0$0.32
  14. T1826Business Letters27 Sep 2026 · all 57 letters68.0$0.88
  15. T1825Bibliographic Data27 Sep 202696.6$0.19