GPT-5.2 (medium)
on τ-bench (airline, clean)
67.9%
Accuracy
0.82
Overall Reliability
#7
of 15 agents
0.76
Consistency
0.74
Predictability
0.95
Robustness
0.94
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.764 |
Outc
0.692
Traj-D
0.844
Traj-S
0.767
Res
0.793
|
| Predictability | 0.735 |
Cal
0.768
AUROC
0.541
Brier
0.735
|
| Robustness | 0.950 |
Fault
1.000
Struct
0.849
Prompt
1.000
|
| Safety | 0.936 |
Harm
0.444
Comp
0.885
Safety
0.936
|
Per-Task Outcome Consistency
Each cell represents a task. Color shows outcome consistency across runs. Hover to see task ID.
Consistency Distribution
KDE of per-task outcome consistency. Peaks at 0 or 1 indicate polarized behavior.
Per-Task Cost Distribution
KDE of mean cost per task (averaged across runs).
Per-Task Time Distribution
KDE of mean execution time per task (averaged across runs).
Calibration Curve
Risk-Coverage (AURC)
Confidence Distribution
Distribution of expressed confidence values across tasks.