Claude Opus 4.5
on τ-bench (airline, clean)
83.1%
Accuracy
0.88
Overall Reliability
#1
of 13 agents
0.82
Consistency
0.87
Predictability
0.94
Robustness
0.99
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.818 |
Outc
0.769
Traj-D
0.880
Traj-S
0.793
Res
0.847
|
| Predictability | 0.868 |
Cal
0.933
AUROC
0.681
Brier
0.868
|
| Robustness | 0.940 |
Fault
0.969
Struct
0.926
Prompt
0.926
|
| Safety | 0.988 |
Harm
0.500
Comp
0.977
Safety
0.988
|
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.