University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

mistral-small-2603

Mistral
average±3.5 (95%)
59.6
leaderboard rankof 114
≈92
benchmarksof 12
12
per 1,000 documentsmean over its benchmarks
$0.87
seconds per inputmean response time
5.1
Compare with other models Its row in the Matrix All its test runs

By task

  • Data correction

    44.3

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

  • Document understanding

    55.1

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

  • Information extraction

    73.8

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

  • Transcription

    59.9

    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 Dataother prompt 94.6±1.9 ≈57of 115 $1.39$0.007 run 7.25 T1942 · 24 Oct 2026prompt_v2.txt
Blacklist Cards 76.1±6.0 ≈106of 114 $0.50$0.016 run 2.01 T1949 · 24 Oct 2026prompt.txt
Book Advert XML Repairother prompt 5.5±3.0 ≈105of 119 $0.64$0.032 run 4.84 T1951 · 24 Oct 2026prompt_v2.txt
Business Letters 46.0±5.5 ≈107of 114 $0.71$0.040 run 2.14 T1943 · 24 Oct 2026prompt.txt
Company Listsother prompt 73.3±9.8 ≈58of 114 $0.84$0.013 run 4.83 T1950 · 24 Oct 2026prompt_v2.txt
Duty Rosters 68.6±4.1 ≈78of 107 $1.44$0.030 run 7.68 T1955 · 24 Oct 2026prompt.txt
Fraktur Adverts 63.6±33.2 ≈67of 114 $1.55$0.008 run 14.97 T1946 · 24 Oct 2026prompt_optimized.txt
General Meeting Minutes 76.8±13.2 ≈25of 108 $0.91$0.008 run 5.07 T1954 · 24 Oct 2026prompt.txt
Library Cardsother prompt 81.5±1.6 ≈81of 114 $0.39$0.10 run 1.48 T1947 · 24 Oct 2026prompt_year-string.txt
Magazine Pagesother prompt 13.6±9.3 ≈66of 115 $0.44$0.020 run 1.86 T1953 · 24 Oct 2026prompt_explicit-json.txt
Medieval Manuscripts 33.0±13.6 ≈95of 114 $0.53$0.006 run 3.61 T1948 · 24 Oct 2026prompt.txt
Personnel Cards 83.1±3.5 ≈94of 114 $1.15$0.070 run 5.13 T1952 · 24 Oct 2026prompt.txt

Price

FromInput, $ per million tokensOutput, $ per million tokensSource
24 Oct 20260.150.6archived price page

Every run

  1. T1955Duty Rosters24 Oct 202668.6$0.030
  2. T1954General Meeting Minutes24 Oct 202676.8$0.008
  3. T1953Magazine Pages24 Oct 202613.6$0.020
  4. T1952Personnel Cards24 Oct 202683.1$0.070
  5. T1951Book Advert XML Repair24 Oct 20265.5$0.032
  6. T1950Company Lists24 Oct 202673.3$0.013
  7. T1949Blacklist Cards24 Oct 202676.1$0.016
  8. T1948Medieval Manuscripts24 Oct 202633.0$0.006
  9. T1947Library Cards24 Oct 202681.5$0.10
  10. T1946Fraktur Adverts24 Oct 202663.6$0.008
  11. T1945Business Letters24 Oct 2026 · 32 signed letters50.0$0.025
  12. T1944Business Letters24 Oct 2026 · 25 unsigned letters32.0$0.015
  13. T1943Business Letters24 Oct 2026 · all 57 letters46.0$0.040
  14. T1942Bibliographic Data24 Oct 202694.6$0.007