Model selection guides
Learn how to read the rankings, compare reasoning efforts and assess a model upgrade for your work.
Compare the models you care about
Compare models and reasoning efforts
Put all three scores, evaluation time and reference cost side by side.
Choose models to compare →Backend & testing0–100Frontend & interaction0–100Knowledge & reasoning0–100
Comparisons from current results
Selection and usage guides
Choosing a reasoning effort
Compare scores, evaluation time and reference cost at neighboring efforts.
Balancing quality, time and cost
Start with an acceptable quality level, then explore faster or cheaper configurations.
Using a model ranking
Separate overall performance, task strengths and tested configurations.
Upgrade differences or regressions?
Compare generations first; a regression requires historical evidence for the same configuration.
Model results directory
Explore backend, frontend and reasoning scores at each tested effort.
Browse all model results