University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

moonshotai/kimi-k3

Moonshot AIvia OpenRouter
average±1.5 (95%)
87.7
leaderboard rankof 114
≈1
benchmarksof 12
12
per 1,000 documentsmean over its benchmarks
$133
seconds per inputmean response time
204.5
Compare with other models Its row in the Matrix All its test runs

By task

  • Data correction

    78.5

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

  • Document understanding

    93.5

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

  • Information extraction

    88.7

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

  • Transcription

    92.4

    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 98.3±1.4 ≈1of 115 $115$0.58 run 198.78 T1390 · 25 Jul 2026prompt.txt
Blacklist Cards 94.8±3.0 ≈14of 114 $32$1.04 run 44.94 T1397 · 25 Jul 2026prompt.txt
Book Advert XML Repair 59.6±9.9 ≈1of 119 $76$3.79 run 133.55 T1400 · 25 Jul 2026prompt.txt
Business Letters 63.0±6.3 ≈24of 114 $53$3.05 run 58.84 T1391 · 17 Sep 2026prompt.txt
Company Lists 94.9±2.9 ≈6of 114 $69$1.03 run 48.35 T1398 · 17 Sep 2026prompt.txt
Duty Rosters 88.7±3.8 ≈10of 107 $317$6.66 run 281.93 T1404 · 18 Sep 2026prompt.txt
Fraktur Adverts 98.4±0.2 ≈9of 114 $300$1.50 run 554.45 T1394 · 17 Sep 2026prompt_optimized.txt
General Meeting Minutes 90.1±7.1 ≈1of 108 $166$1.50 run 298.12 T1403 · 25 Jul 2026prompt.txt
Library Cards 91.2±1.4 ≈1of 114 $43$11 run 40.63 T1395 · 17 Sep 2026prompt.txt
Magazine Pages 94.6±5.0 ≈1of 115 $70$3.22 run 144.0 T1402 · 25 Jul 2026prompt.txt
Medieval Manuscripts 81.6±9.7 ≈27of 114 $268$3.22 run 519.97 T1396 · 25 Jul 2026prompt.txt
Personnel Cards 97.3±1.0 ≈1of 114 $88$5.38 run 130.41 T1401 · 25 Jul 2026prompt.txt

Price

FromInput, $ per million tokensOutput, $ per million tokensSource
16 Sep 20262.64813813.282724archived price page
25 Jul 2026315archived price page

Every run

  1. T1404Duty Rosters18 Sep 202688.7$6.66
  2. T1403General Meeting Minutes18 Sep 202682.7$0.98
  3. T1402Magazine Pages18 Sep 202685.9$2.60
  4. T1401Personnel Cards17 Sep 202696.1$4.55
  5. T1400Book Advert XML Repair17 Sep 202647.3$2.86
  6. T1399Company Lists17 Sep 202681.9$0.81
  7. T1398Company Lists17 Sep 202694.9$1.03
  8. T1397Blacklist Cards17 Sep 202691.5$1.11
  9. T1396Medieval Manuscripts17 Sep 202639.5$1.91
  10. T1395Library Cards17 Sep 202691.2$11
  11. T1394Fraktur Adverts17 Sep 202698.4$1.50
  12. T1393Business Letters17 Sep 2026 · 32 signed letters64.0$2.06
  13. T1392Business Letters17 Sep 2026 · 25 unsigned letters65.0$1.28
  14. T1391Business Letters17 Sep 2026 · all 57 letters63.0$3.05
  15. T1390Bibliographic Data17 Sep 202698.0$0.65
  16. T1404Duty Rosters25 Jul 202688.1$6.17
  17. T1403General Meeting Minutes25 Jul 202690.1$1.50
  18. T1402Magazine Pages25 Jul 202694.6$3.22
  19. T1401Personnel Cards25 Jul 202697.3$5.38
  20. T1400Book Advert XML Repair25 Jul 202659.6$3.79
  21. T1399Company Lists25 Jul 202676.5$0.89
  22. T1398Company Lists25 Jul 202687.8$1.11
  23. T1397Blacklist Cards25 Jul 202694.8$1.04
  24. T1396Medieval Manuscripts25 Jul 202681.6$3.22
  25. T1395Library Cards25 Jul 202688.7$14
  26. T1394Fraktur Adverts25 Jul 202696.4$1.63
  27. T1393Business Letters25 Jul 2026 · 32 signed letters65.0$1.96
  28. T1392Business Letters25 Jul 2026 · 25 unsigned letters62.0$1.31
  29. T1391Business Letters25 Jul 2026 · all 57 letters59.0$3.22
  30. T1390Bibliographic Data25 Jul 202698.3$0.58