PUBLIC DATA
API & benchmark explorer
Compare provider routes and benchmark results with their original evaluation settings.
Compare scores within the same evaluation. Reasoning effort, tools and agent settings can change the result.
Scatter plot
X
Y
Bar chart · top 15
Metric
Source health
Fetch time and evaluation date are different. Failed updates retain the last valid snapshot.
Further evaluation sources
These sources need additional verification or reproducible evaluations before their scores can be included.
Each source retains its own license and attribution requirements. Missing values are not zero. Catalog listings do not guarantee current availability.