University of Basel RISE · Research and Infrastructure SupportHumanities Data Benchmark

gemini-exp-1206

Google

legacy model Its test configurations are retired, so it is no longer run. The figures are from its best legacy runs, scored as they ran at the time; it is not ranked.

average
77.2
benchmarksof 12
2
seconds per inputmean response time
94.8
Compare with other models Its row in the Matrix All its test runs

By benchmark

BenchmarkScoreThe best legacy 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. Legacy models are not ranked.Per 1,000 docss / inputBest run
Business Lettersolder run 58.0 – –– run 14.54 T0014 · 1 Apr 2025prompt.txt
Fraktur Advertsolder run 96.3 – –– run 175.14 T0087 · 9 May 2025prompt_optimized.txt

Not run on: Bibliographic Data, Blacklist Cards, Book Advert XML Repair, Company Lists, Duty Rosters, General Meeting Minutes, Library Cards, Magazine Pages, Medieval Manuscripts, Personnel Cards.

Every run

  1. T0087Fraktur Adverts9 May 2025 · legacy configuration96.3–
  2. T0087Fraktur Adverts7 May 2025 · legacy configuration89.6–
  3. T0014Business Letters11 Apr 2025 · all 57 letters · legacy configuration58.0–
  4. T0014Business Letters1 Apr 2025 · all 57 letters · legacy configuration58.0–
  5. T0014Business Letters5 Mar 2025 · all 57 letters · legacy configuration50.0–