external benchmark
Vals SWE-bench mirror leaderboard
Vals-hosted SWE-bench mirror. Every model the catalog carries a published Vals SWE-bench mirror value for, ranked by that value.
Categoryexternal
MeasureAccuracy score
TasksSoftware engineering issue-resolution tasks
DifficultyProduction software engineering
Published byVals SWE-bench
Vals AI hosted SWE-bench view for solving production software engineering tasks.
No model in the catalog has a published Vals SWE-bench mirror score.
The benchmark is defined by Vals SWE-bench, but the catalog carries no value for it yet. An absent value is shown as absent here rather than as a zero.