Model selection guides
Learn how to read the rankings, choose a reasoning effort, and evaluate model upgrades for your work.
Compare models and reasoning efforts
Put dimension scores, evaluation time, and reference cost side by side.
Choose models to compareComparisons from current results
Model and effort comparisons
GPT-6.1 Sol vs Claude Opus 5.5
Compare each model at its highest-scoring tested overall effort. Start with the dimension relevant to your task, then weigh evaluation time and reference cost.
GPT-6.1 Sol vs GPT-6 Sol
Compare the two Sol generations at their highest-scoring tested overall efforts, and see how score deltas trade off against evaluation time and reference cost.
GLM 5.3 High vs Max
Keep the GLM 5.3 model fixed and compare its tested High and Max reasoning efforts side by side.
Selection and usage guides
Choosing a reasoning effort
Compare capability scores, evaluation time, and reference cost across neighboring efforts.
Balancing quality, time, and cost
Start with an acceptable quality threshold, then explore faster or lower-cost configurations.
Using a model ranking
Separate overall scores, dimension strengths, and tested configurations.
Upgrade differences or regressions?
Compare generations first; confirming a regression requires historical results for the exact same configuration.
Model results directory
Explore scores and subscores across all tested efforts.
Browse all model results