Agent: GPT-4o Mini

Reliability Dimensions by Benchmark
GAIA Detail →
Accuracy 26.3%
Consistency 0.73
Predictability 0.58
Robustness 0.97
Safety 1.00
Reliability 0.76
τ-bench (airline, clean) Detail →
Accuracy 29.5%
Consistency 0.74
Predictability 0.35
Robustness 0.93
Safety 0.85
Reliability 0.67
τ-bench (airline, original) Detail →
Accuracy 21.3%
Consistency 0.76
Predictability 0.32
Robustness 0.92
Safety 0.76
Reliability 0.67