Why are almost all new benchmarks and leaderboards coding focused?

I know in in this community LLM's are generally used for coding but there are other usecases besides coding and those usecases should be tested too. I also know benchmarks can sometimes be benchmaxxed and the model...