arXiv:2607. 17082v1 Announce Type: new Abstract: Large language model agents solve tasks by generating trajectories that interleave planning, tool calls, and intermediate results.
arXiv:2607. 17082v2 Announce Type: replace-cross Abstract: Large language model agents solve tasks by generating trajectories that interleave planning, tool calls, and intermediate results.
GRASP is a multi-stage planning framework that improves the reliability of large language models on complex tasks. It separates planning into three specialized modules—GenPlan for global macro-guidelines, RevPlan for exploring localized strategies, and VerPlan for multi-criteria evaluation—allowing context isolation and strict macro-regularization. Experiments show GRASP outperforms direct LLM planners by significant margins on datasets such as Natural Plan Calendar Scheduling, ZebraLogic, and SciBench Math, and it mitigates performance collapse in multi-task and dual-task settings.
LLM‑BabyBench transforms the BabyAI gridworld into a fully observable, purely textual setting that isolates planning as the sole source of failure. By serialising the entire grid, providing formal instructions, and validating actions deterministically, the benchmark introduces the PPD suite—Predict, Plan, and Decompose tasks—each scored with metrics that separate mission understanding from sequencing. Across a range of large language models, simulation accuracy is high while planning success drops sharply beyond a model‑specific horizon, revealing that plan length—not grid size—drives failure and that models often commit to a single corridor‑shaped route without backtracking.
Translating natural-language planning intent into verified plans is a longstanding challenge: people communicate goals in language, while classical planners require formal PDDL specifications. Recent agentic frameworks bridge this gap by orchestrating a pool of specialized repair agents inside a verifier-checked refinement loop, but the orchestrator at the centre is itself a prompted frontier LLM, paying a frontier-LLM API call at every refinement step.
GRASP is a multi-stage planning framework that separates planning into specialized modules: GenPlan for global macro-guidelines, RevPlan for exploring localized strategies, and VerPlan for multi-criteria evaluation. This strategy-aware approach yields state‑of‑the‑art accuracy on diverse datasets, outperforming direct LLM planners by up to 30.8% on ZebraLogic and reducing multi‑task degradation. GRASP’s context isolation and macro‑regularization also give it a 14.5% edge over frontier reasoning models like GPT‑5‑mini.
The paper introduces complete cyclic subtask graphs for large language model agents, enabling a workflow controller where all subtasks are fully connected and a unified agent selects transitions based on natural‑language criteria. It evaluates task‑specific and benchmark‑generic cyclic graphs on TextCraft, ALFWorld, and Finance‑Agent, comparing them to ReAct and dependency‑directed workflows, and identifies three distinct workflow signatures that influence the effectiveness of cyclic routing. The study also provides a workflow‑signature matrix, robustness analysis, token‑cost accounting, and failure‑mode structure, concluding that cyclic subtask graphs serve as a diagnostic tool to determine when flexible backtracking is worthwhile versus when simpler controllers suffice.
arXiv:2508. 02721v2 Announce Type: replace-cross Abstract: While powerful, the inherent non-determinism of large language model (LLM) agents limits their application in structured operational environments where procedural fidelity and predictable execution are strict requirements.
arXiv:2609.22813v1 Announce Type: cross Abstract: We present \emph{commonsense ranked search} (CoRS), a novel path planner that turns an abstract instruction into a route that follows commonsense. Wh...
The paper introduces InFlowOp, a label‑free optimization framework that assigns costs to each decision in a multi‑agent workflow, balancing agent competence against execution time. It determines task granularity and agent assignment before execution and corrects faults during execution using the same cost metric. The authors also present Braid, a benchmark for multi‑agent coordination, and show that InFlowOp outperforms single‑agent baselines by up to 11.97% across various domains.
arXiv:2605. 12925v3 Announce Type: replace-cross Abstract: Evaluation of software engineering (SWE) agents is dominated by a binary signal: whether the final patch passes the tests.
arXiv:2606. 27806v1 Announce Type: new Abstract: World models for language agents come in two useful forms.
arXiv:2606. 17454v1 Announce Type: new Abstract: AI agent performance is not just a modeling problem, it is fundamentally a systems problem.