GPT-5.2 (xhigh)
on τ-bench (airline, clean)
67.7%
Accuracy
0.81
Overall Reliability
#4
of 13 agents
0.70
Consistency
0.78
Predictability
0.96
Robustness
0.95
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.698 |
Outc
0.538
Traj-D
0.845
Traj-S
0.731
Res
0.768
|
| Predictability | 0.776 |
Cal
0.810
AUROC
0.747
Brier
0.776
|
| Robustness | 0.965 |
Fault
1.000
Struct
1.000
Prompt
0.894
|
| Safety | 0.954 |
Harm
0.400
Comp
0.923
Safety
0.954
|
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.