GraphPlanner: Graph Memory-Augmented Agentic Routing for Multi-Agent LLMs

Tao Feng, Haozhen Zhang, Zijie Lei, Peixuan Han, Jiaxuan You · 2026

Read GraphPlanner: Graph Memory-Augmented Agentic Routing for Multi-Agent LLMs by Tao Feng, Haozhen Zhang, Zijie Lei, Peixuan Han, Jiaxuan You (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-23626.txt.

Description

Feng and colleagues present GraphPlanner, a router for multi-agent language-model systems that builds a workflow per query from a memory of past queries, agents, and responses. Each step picks both the model backbone and the agent role, such as planner, executor, or summariser, under reinforcement learning. Across fourteen tasks it lifts accuracy by about 9 percent while slashing GPU memory from over 180 GiB to about 1 GiB.

Catalog record

CREATORTao Feng, Haozhen Zhang, Zijie Lei, Peixuan Han, Jiaxuan You
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE95465
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.23626

Browse the full library · About the Tower of Babel