做数学推理或智能体开发的团队值得关注——LEAP 用通用模型+验证反馈循环就超越了专用系统，说明智能体框架设计比模型本身更关键，建议点开论文看具体架构。
Google 发布新研究 LEAP（Lean-Enhanced Agentic Programming），通过将通用大语言模型封装在智能体框架中，每一步都基于 Lean 编译器进行验证，并迭代利用验证器反馈。该框架使同一个通用模型解决了全部 12 道 Putnam 2025 数学竞赛题，并将 Lean-IMO-Bench 的一次性求解率从不到 10% 提升至 70%，超越了得分为 48% 的专用金牌系统。这项研究展示了定制智能体框架在数学推理任务上的巨大潜力，论文已发布在 arXiv 上。
New research from Google. Just shows the impressive results you can get from custom agent harnesses...
New research from Google. Just shows the impressive results you can get from custom agent harnesses. LEAP wraps a general-purpose LLM in an agentic scaffold that grounds every step in the Lean compiler and iterates against verifier feedback. The same general model solves all 12 Putnam 2025 problems and lifts Lean-IMO-Bench one-shot solve rate from under 10% to 70%, beating a specialized gold-medal system that scores 48%. Paper: arxiv.org/abs/2606.03303 Learn to build effective AI agents in our academy: academy.dair.ai 💬 2 🔄 4 ❤️ 17 👀 1537 📊 9 ⚡