Number of consents.
AI agent sometimes asks user to consent to the function calling, and perform it at the next step.
This property represents the number of consents to allow. If the number of consents from the AI agent exceeds the configured value, the benchmark will be failed.
Repeat count.
The number of repeating count for the benchmark execution for each scenario.
Simultaneous count.
The number of simultaneous count for the parallel benchmark execution.
If you configure this property greater than 1, the benchmark
for each scenario would be executed in parallel in the given
count.
Configuration for the benchmark.
AgenticaSelectBenchmark.IConfigis a data structure which represents a configuration for the benchmark, especially the capacity information of the benchmark execution.