Pass@1 i
—
No completed experiment
Evaluation workspace
Know exactly how your agent arrived.
Run repeatable chart tasks, inspect every tool decision, and grade outcomes with evidence—not vibes.
Pass@k i
—
No completed experiment
Tool-path precision i
—
0 calls judged
Median latency
—
P95 —
Latest activity
Experiments
No experiments yet
Create an experiment to generate repeated candidate runs and grades.
Create experiment