> VERY verbose in the plan output

Is that an issue? GitHub charges per-request, not per-token, so a verbose output and short output will be the same cost

What model are you using?