Legacy Integration Bench
Trajectories are published here when the first public task set ships.
What is plannedtrajectories
Detailed published attempts across the benchmarks: every tool call, vendor request, graded check and diff.
Live · rev 01
Open the pickerIntegration Bench
1050 published attempts · 21 models · 50 public tasks.
no verdict
checks passed 100%0%
In authoring · Late 2026