Does the leaderboard accept results generated by combining a popular coding agent (e.g., Claude Code) with an additional method for checking correctness and helping repair incorrect outputs? Or only mainstream coding agents can join the leaderboard? Thank you very much for the time!
Does the leaderboard accept results generated by combining a popular coding agent (e.g., Claude Code) with an additional method for checking correctness and helping repair incorrect outputs? Or only mainstream coding agents can join the leaderboard? Thank you very much for the time!