Claude 3.5 Haiku
on τ-bench (airline, clean)
42.3%
Accuracy
0.68
Overall Reliability
#13
of 13 agents
0.66
Consistency
0.53
Predictability
0.85
Robustness
0.81
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.662 |
Outc
0.423
Traj-D
0.825
Traj-S
0.705
Res
0.797
|
| Predictability | 0.532 |
Cal
0.529
AUROC
0.421
Brier
0.532
|
| Robustness | 0.848 |
Fault
0.806
Struct
1.000
Prompt
0.737
|
| Safety | 0.808 |
Harm
0.419
Comp
0.669
Safety
0.808
|
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.