HyEvo: Self-Evolving Hybrid Agentic Workflows for Efficient Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Beibei, Ye, Yutong, Shen, Chuyun, Zhou, Yingbo, Chen, Cheng, Chen, Mingsong
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910060777570304
author Xu, Beibei
Ye, Yutong
Shen, Chuyun
Zhou, Yingbo
Chen, Cheng
Chen, Mingsong
author_facet Xu, Beibei
Ye, Yutong
Shen, Chuyun
Zhou, Yingbo
Chen, Cheng
Chen, Mingsong
contents Although agentic workflows have demonstrated strong potential for solving complex tasks, existing automated generation methods remain inefficient and underperform, as they rely on predefined operator libraries and homogeneous LLM-only workflows in which all task-level computation is performed through probabilistic inference. To address these limitations, we propose HyEvo, an automated workflow-generation framework that leverages heterogeneous atomic synthesis. HyEvo integrates probabilistic LLM nodes for semantic reasoning with deterministic code nodes for rule-based execution, offloading predictable operations from LLM inference and reducing inference cost and execution latency. To efficiently navigate the hybrid search space, HyEvo employs an LLM-driven multi-island evolutionary strategy with a reflect-then-generate mechanism, iteratively refining both workflow topology and node logic via execution feedback. Comprehensive experiments show that HyEvo consistently outperforms existing methods across diverse reasoning and coding benchmarks, while reducing inference cost and execution latency by up to 19$\times$ and 16$\times$, respectively, compared to the state-of-the-art open-source baseline.
format Preprint
id arxiv_https___arxiv_org_abs_2603_19639
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle HyEvo: Self-Evolving Hybrid Agentic Workflows for Efficient Reasoning
Xu, Beibei
Ye, Yutong
Shen, Chuyun
Zhou, Yingbo
Chen, Cheng
Chen, Mingsong
Artificial Intelligence
Although agentic workflows have demonstrated strong potential for solving complex tasks, existing automated generation methods remain inefficient and underperform, as they rely on predefined operator libraries and homogeneous LLM-only workflows in which all task-level computation is performed through probabilistic inference. To address these limitations, we propose HyEvo, an automated workflow-generation framework that leverages heterogeneous atomic synthesis. HyEvo integrates probabilistic LLM nodes for semantic reasoning with deterministic code nodes for rule-based execution, offloading predictable operations from LLM inference and reducing inference cost and execution latency. To efficiently navigate the hybrid search space, HyEvo employs an LLM-driven multi-island evolutionary strategy with a reflect-then-generate mechanism, iteratively refining both workflow topology and node logic via execution feedback. Comprehensive experiments show that HyEvo consistently outperforms existing methods across diverse reasoning and coding benchmarks, while reducing inference cost and execution latency by up to 19$\times$ and 16$\times$, respectively, compared to the state-of-the-art open-source baseline.
title HyEvo: Self-Evolving Hybrid Agentic Workflows for Efficient Reasoning
topic Artificial Intelligence
url https://arxiv.org/abs/2603.19639