Clarification on "Overall Score" calculation logic

[yikeyang33]

Hi team,

I have successfully ran the benchmark on my local model. However, I noticed that the final "Overall Score" (the metric used in the Leaderboard) is not explicitly saved in the generated benchmark_results_*.json file.

The JSON file contains detailed metrics for each task file (e.g., task_completion_score, tool_selection_score, etc.), but I couldn't find the aggregated final score or the explicit formula in the codebase/documentation to derive it.

Could you please clarify the exact formula used to calculate the Overall Score from these JSON metrics?

Thanks

[yikeyang33]

I realized that #12 #14 has already answered this question, so I’m closing this issue.