alanhuangyoo/crux

Taking the pi coding agent to Claude Code-level performance: 0.539 → 0.773 pass@1 on Terminal-Bench 2.1 with the same self-hosted 27B model, by re-engineering the agent.

语言：TypeScript · GitHub 星标：186 · Fork：36

话题：agent-evaluation · ai-agent · coding-agent · context-engineering · llm · terminal-bench

在 GitHub 上查看原仓库