University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

thinkingmachines/Inkling:together

Thinking Machinesvia Hugging Face
average±3.1 (95%)
34.1
leaderboard rankof 114
113
benchmarksof 12
12
per 1,000 documentsmean over its benchmarks
$10
seconds per inputmean response time
39.5
Compare with other models Its row in the Matrix All its test runs

By task

  • Data correction

    20.3

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

  • Document understanding

    12.2

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

  • Information extraction

    50.0

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

  • Transcription

    9.4

    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 83.2±11.8 ≈74of 115 $26$0.13 run 111.08 T1555 · 17 Sep 2026prompt.txt
Blacklist Cards 90.4±3.3 ≈51of 114 $5.16$0.17 run 11.45 T1562 · 15 Aug 2026prompt.txt
Book Advert XML Repair 17.6±8.7 ≈64of 119 $8.30$0.42 run 34.06 T1565 · 15 Aug 2026prompt.txt
Business Letters 53.0±9.0 ≈90of 114 $8.87$0.43 run 23.45 T1556 · 15 Aug 2026prompt.txt
Company Listsother prompt 33.7±19.4 ≈105of 114 $8.22$0.12 run 28.87 T1564 · 15 Aug 2026prompt_min.txt
Duty Rosters 0.0±0.0 ≈105of 107 $12$0.24 run 50.28 T1569 · 15 Aug 2026prompt.txt
Fraktur Adverts 0.0±0.0 ≈108of 114 $11$0.055 run 41.66 T1559 · 15 Aug 2026prompt_optimized.txt
General Meeting Minutes 9.3±18.2 ≈107of 108 $11$0.098 run 45.09 T1568 · 15 Aug 2026prompt.txt
Library Cards 80.7±1.4 ≈90of 114 $6.58$1.73 run 36.12 T1560 · 18 Sep 2026prompt.txt
Magazine Pages 13.5±9.8 ≈66of 115 $5.06$0.23 run 12.26 T1567 · 15 Aug 2026prompt.txt
Medieval Manuscripts 5.2±10.1 ≈112of 114 $11$0.13 run 42.52 T1561 · 16 Aug 2026prompt.txt
Personnel Cards 23.1±12.0 ≈111of 114 $11$0.66 run 36.73 T1566 · 15 Aug 2026prompt.txt

Price

FromInput, $ per million tokensOutput, $ per million tokensSource
15 Aug 20261.04.05archived price page

Every run

  1. T1567Magazine Pages18 Sep 202612.3$0.30
  2. T1562Blacklist Cards18 Sep 202689.3$0.20
  3. T1560Library Cards18 Sep 202680.7$1.73
  4. T1559Fraktur Adverts17 Sep 2026––
  5. T1558Business Letters17 Sep 2026 · 32 signed letters56.0$0.33
  6. T1557Business Letters17 Sep 2026 · 25 unsigned letters46.0$0.26
  7. T1556Business Letters17 Sep 2026 · all 57 letters49.0$0.61
  8. T1555Bibliographic Data17 Sep 202683.2$0.13
  9. T1561Medieval Manuscripts16 Aug 20265.2$0.13
  10. T1569Duty Rosters15 Aug 20260.0$0.24
  11. T1568General Meeting Minutes15 Aug 20269.3$0.098
  12. T1567Magazine Pages15 Aug 202613.5$0.23
  13. T1566Personnel Cards15 Aug 202623.1$0.66
  14. T1565Book Advert XML Repair15 Aug 202617.6$0.42
  15. T1564Company Lists15 Aug 202633.7$0.12
  16. T1563Company Lists15 Aug 202626.4$0.15
  17. T1562Blacklist Cards15 Aug 202690.4$0.17
  18. T1560Library Cards15 Aug 202670.9$1.72
  19. T1559Fraktur Adverts15 Aug 20260.0$0.055
  20. T1558Business Letters15 Aug 2026 · 32 signed letters58.0$0.23
  21. T1557Business Letters15 Aug 2026 · 25 unsigned letters51.0$0.17
  22. T1556Business Letters15 Aug 2026 · all 57 letters53.0$0.43
  23. T1555Bibliographic Data15 Aug 20260.0$0.055