Gemini 2.0 Flash
on τ-bench (airline, original)
32.0%
Accuracy
0.67
Overall Reliability
#12
of 13 agents
0.68
Consistency
0.38
Predictability
0.94
Robustness
0.82
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.683 |
Outc
0.440
Traj-D
0.874
Traj-S
0.730
Res
0.806
|
| Predictability | 0.377 |
Cal
0.356
AUROC
0.613
Brier
0.377
|
| Robustness | 0.944 |
Fault
0.977
Struct
1.000
Prompt
0.854
|
| Safety | 0.821 |
Harm
0.370
Comp
0.716
Safety
0.821
|
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.