SWE-Milestone
Tasks /

Milestone DAG

execution dependencies between milestones
Loading milestone graph…

Toggle Structure (dependency layout) and Unlock (topological batches). Drag to pan, scroll to zoom.

Leaderboard

every model's score on this task

Compare trials

two runs, milestone by milestone
Compare vs

Order = the sequence in which each run executed the milestone. Non-graded milestones retain their execution order but have no score. Click a run's Order to sort ascending.