AGENTIC-SYSTEMSJuly 13, 20261 min read

GATS: Graph-Augmented Tree Search with Layered World Models for Efficient Agent Planning

Researchers have developed GATS, a breakthrough planning framework for AI agents that completely eliminates LLM calls during inference. Achieving a 100% success rate in complex tasks, GATS paves the way for highly efficient and deterministic autonomous agents.

Computer Science > Artificial Intelligence

Title:GATS: Graph-Augmented Tree Search with Layered World Models for Efficient Agent Planning

View PDF HTML (experimental)Abstract:Large Language Model (LLM) agents have shown promise in multi-step planning tasks, but existing approaches like LATS (Language Agent Tree Search) and ReAct rely heavily on LLM inference during planning, leading to high computational costs and stochastic behavior. We present \textbf{GATS} (Graph-Augmented Tree Search), a planning framework that combines systematic UCB1-based tree search with a layered world model to eliminate LLM calls during inference while achieving superior planning performance. Our three-layer world model integrates: (L1) exact symbolic action matching, (L2) statistics learned from execution logs, and (L3) LLM-based prediction for unknown actions. On synthetic planning tasks with branching paths and dead-ends, GATS achieves \textbf{100% success rate} compared to 92 % for LATS and 64% for ReAct. On a comprehensive stress test spanning 12 challenging scenarios -- including coding workflows, web navigation, and long-horizon tasks -- GATS maintains \textbf{100% success} while LATS drops to 88.9 % and ReAct to 23.9%. GATS requires \textbf{zero LLM calls per task} during planning (vs. 37 per task for LATS) and produces deterministic plans with zero variance across runs. Our results demonstrate that systematic search with learned world models can substantially outperform LLM-guided exploration for agent planning.

Bibliographic and Citation Tools

Code, Data and Media Associated with this Article

Demos

Recommenders and Search Tools

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Source: arXiv cs.AI Recent

More in this category

agentic-systems

Interval Certifications for Multilayered Perceptrons via Lattice Traversal

Researchers have proposed a rigorous theoretical framework for AI safety by reducing the adversarial robustness problem to a lattice traversal problem, introducing a novel 'complete certification' alongside traditional 'sound certification'.

agentic-systems

Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading

Researchers have introduced Long-Horizon-Terminal-Bench, a new benchmark designed to test the limits of AI agents on complex, long-horizon tasks. Featuring a dense reward-based grading system, the benchmark reveals that even the most advanced frontier models struggle with tasks requiring hours of execution and iterative debugging.

agentic-systems

Context Graphs for Proactive Enterprise Agents

Researchers propose 'Context Graphs' to enable proactive AI agents in enterprises, shifting from reactive query-response to active information delivery. The system reduces information delivery time from 47 minutes to under 30 seconds with high precision.

NOW LET US Related – Evaluating the Effect of Frame Rate in Sequence-Based Classification of Autism-Related Self-Stimulatory Hand Idiosyncrasies

agentic-systems

Evaluating the Effect of Frame Rate in Sequence-Based Classification of Autism-Related Self-Stimulatory Hand Idiosyncrasies

A new study demonstrates that optimizing frame sampling rates and neural network architectures can significantly improve the accuracy of detecting autism-related self-stimulatory behaviors from video, achieving up to 98.75% accuracy using GRU models at a 15-frame interval.

agentic-systems

From Atomic Actions to Standard Operating Procedures: Iterative Tool Optimization for Self-Evolving LLM Agents

Researchers have introduced EvoSOP, a framework that enables LLM agents to synthesize granular atomic actions into reusable Standard Operating Procedures (SOPs). This iterative tool optimization significantly boosts task success rates and reduces reasoning overhead, paving the way for self-evolving AI systems.

agentic-systems

Evaluating SageMath-Augmented LLM Agents for Computational and Experimental Mathematics

A new study evaluates the integration of SageMath, a Computer Algebra System (CAS), into LLM agents to solve research-level mathematical problems. The results show significant performance gains across all models, with GPT-5.5 achieving the highest solve rate of 75.2%, paving the way for automated conjecture discovery.