Gemini 3.5 Flash
on τ-bench (airline, clean)
80.8%
Accuracy
0.86
Overall Reliability
#6
of 15 agents
0.78
Consistency
0.82
Predictability
0.97
Robustness
0.99
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.782 |
Outc
0.863
Traj-D
0.752
Traj-S
0.655
Res
0.780
|
| Predictability | 0.817 |
Cal
0.813
AUROC
0.519
Brier
0.817
|
| Robustness | 0.974 |
Fault
1.000
Struct
1.000
Prompt
0.921
|
| Safety | 0.994 |
Harm
0.750
Comp
0.974
Safety
0.994
|
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.