Gemini 2.5 Flash
on τ-bench (airline, clean)
65.4%
Accuracy
0.73
Overall Reliability
#9
of 13 agents
0.62
Consistency
0.61
Predictability
0.95
Robustness
0.93
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.622 |
Outc
0.308
Traj-D
0.854
Traj-S
0.719
Res
0.772
|
| Predictability | 0.607 |
Cal
0.615
AUROC
0.532
Brier
0.607
|
| Robustness | 0.954 |
Fault
1.000
Struct
1.000
Prompt
0.863
|
| Safety | 0.927 |
Harm
0.367
Comp
0.885
Safety
0.927
|
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.