Benchmarking pi's harness

[zdenham]

I'm wondering if there have been any efforts to benchmark Pi's harness against others, like codex cli or claude agent sdk.

And if not, I'd like to help! I'm looking to phase out claude agent sdk for my team's internal coding agent, but I need some objective evidence on a diverse / in distribution agentic coding task set that we aren't giving up significant performance out of the box.