CZ-EVAL / leaderboard /klokan.csv
Hynek Kydlicek
new models
acc5a5f
,Elementary 2-3,Elementary 4-5,Elementary 6-7,Elementary 8-9,High School 1-2,High School 3-4
anthropic/claude-2.1,43.96551724137931,50.35971223021583,39.87730061349693,39.75155279503105,33.33333333333333,14.772727272727273
anthropic/claude-3-haiku,37.5,38.405797101449274,27.160493827160494,31.25,32.70440251572327,25.609756097560975
anthropic/claude-3-opus,78.57142857142857,65.94202898550725,61.111111111111114,50.0,45.28301886792453,48.78048780487805
anthropic/claude-3-sonnet,59.82142857142857,60.86956521739131,38.2716049382716,45.0,33.9622641509434,25.609756097560975
cohere/command-r,24.107142857142858,28.26086956521739,29.01234567901235,21.875,24.528301886792452,18.29268292682927
google/gemini-pro,33.92857142857143,28.985507246376812,26.543209876543212,28.125,30.18867924528302,23.170731707317074
google/gemma-7b-it,16.964285714285715,15.942028985507244,16.666666666666664,12.5,20.125786163522015,21.951219512195124
mistralai/mistral-large,51.78571428571429,53.62318840579711,41.358024691358025,37.5,35.22012578616352,23.170731707317074
mistralai/mistral-medium,39.285714285714285,34.78260869565217,28.39506172839506,24.375,27.67295597484277,21.951219512195124
mistralai/mixtral-8x7b-instruct,30.357142857142854,34.05797101449276,28.39506172839506,22.5,22.0125786163522,23.170731707317074
openai/gpt-3.5-turbo,40.17857142857143,39.85507246376812,33.95061728395062,31.874999999999996,26.41509433962264,19.51219512195122
openai/gpt-4-1106-preview,71.42857142857143,69.56521739130434,59.876543209876544,52.5,50.314465408805034,42.68292682926829