GRASP 框架提出分阶段规划方法，多项基准超越 GPT-5-mini
GRASP: Generating, Revising, and Assessing for Strategic Planning with Agentic AI
一篇很硬的论文：把规划拆成生成、修订、评估三个隔离模块，ZebraLogic 上比直接让 LLM 规划高 30.8%，还压过 GPT-5-mini 14.5%。
GRASP 是一种策略感知的多阶段规划框架，将规划流程拆分为 GenPlan、RevPlan、VerPlan 三个上下文隔离的模块。在 Natural Plan Calendar Scheduling 上准确率比直接 LLM 规划高约 12.4%，在 ZebraLogic 上提升约 30.8%。在双任务交错环境中，GRASP 比直接规划取得最高 16.7% 的绝对准确率增益，并消除多任务场景下的性能衰减。通过上下文隔离和宏指令正则化，GRASP 领先 GPT-5-mini 等前沿推理模型 14.5%。
GRASP: Generating, Revising, and Assessing for Strategic Planning with Agentic AI
Large Language Models (LLMs) typically exhibit a performance profile where reliability degrades as task complexity increases. We address the challenge of generating high-quality natural language executable plans for complex tasks by introducing $\textbf{GRASP}$, a strategy-aware, multi-stage planning framework. GRASP decouples the planning pipeline across specialized, context-isolated modules: it pre-compiles global macro-guidelines (GenPlan), explores alternative localized strategies within isolated context windows (RevPlan), and independently evaluates trajectories using a multi-criteria discriminator (VerPlan). Empirical evaluations show that GRASP consistently establishes a new state-of-the-art frontier across diverse datasets, yielding substantial accuracy gains over direct LLM planners on Natural Plan Calendar Scheduling ($\sim$12.4$\%$$\uparrow$), ZebraLogic ($\sim$30.8$\%$$\uparrow$), and SciBench Math. Crucially, under multi-task scaling-where standard planners suffer immediate performance collapse-GRASP completely flattens the multi-task degradation penalty. In interleaved dual-task environments, GRASP achieves an absolute accuracy gain of up to 16.7$\%$ over direct LLM planners. Furthermore, by isolating context and enforcing strict macro-regularization, GRASP outperforms frontier reasoning models (such as GPT-5-mini) by a margin of 14.5$\%$.