Claude Opus 4.5
on τ-bench (airline, clean)
80.8%
Accuracy
0.88
Overall Reliability
#3
of 15 agents
0.84
Consistency
0.85
Predictability
0.97
Robustness
0.98
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.835 |
Outc
0.829
Traj-D
0.859
Traj-S
0.808
Res
0.843
|
| Predictability | 0.848 |
Cal
0.897
AUROC
0.702
Brier
0.848
|
| Robustness | 0.966 |
Fault
1.000
Struct
0.980
Prompt
0.917
|
| Safety | 0.981 |
Harm
0.500
Comp
0.962
Safety
0.981
|
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.