Gemini 3.0 Pro
on τ-bench (airline, original)
58.8%
Accuracy
0.77
Overall Reliability
#3
of 13 agents
0.76
Consistency
0.60
Predictability
0.94
Robustness
0.96
Safety
| Dimension | Agg | Sub-metrics |
|---|---|---|
| Consistency | 0.758 |
Outc
0.620
Traj-D
0.857
Traj-S
0.775
Res
0.837
|
| Predictability | 0.600 |
Cal
0.601
AUROC
0.523
Brier
0.600
|
| Robustness | 0.942 |
Fault
1.000
Struct
0.918
Prompt
0.907
|
| Safety | 0.962 |
Harm
0.208
Comp
0.952
Safety
0.962
|
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.