# FLEX: Continuous Agent Evolution via Forward Learning from Experience

Source: [FLEX: Continuous Agent Evolution via Forward Learning from Experience](<https://www.sota2.com/research/paper/6912c2f82828b7cc6937dadc>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-12-10T05:02:29.550Z

Authors: Zhicheng Cai, Xinyuan Guo, Yu Pei, Jiangtao Feng, Jinsong Su, Jiangjie Chen, Ya-Qin Zhang, Wei-Ying Ma, Mingxuan Wang, Hao Zhou

Published: 2025-11-09T16:31:39.000Z

## Abstract

Autonomous agents driven by Large Language Models (LLMs) have revolutionized reasoning and problem-solving but remain static after training, unable to grow with experience as intelligent beings do during deployment. We introduce Forward Learning with EXperience (FLEX), a gradient-free learning paradigm that enables LLM agents to continuously evolve through accumulated experience. Specifically, FLEX cultivates scalable and inheritable evolution by constructing a structured experience library through continual reflection on successes and failures during interaction with the environment. FLEX delivers substantial improvements on mathematical reasoning, chemical retrosynthesis, and protein fitness prediction (up to 23% on AIME25, 10% on USPTO50k, and 14% on ProteinGym). We further identify a clear scaling law of experiential growth and the phenomenon of experience inheritance across agents, marking a step toward scalable and inheritable continuous agent evolution. Project Page: https://flex-gensi-thuair.github.io.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Clinical Reasoning on MedR-Bench Diagnosis](<https://www.sota2.com/research/sota/clinical-reasoning-on-medr-bench-diagnosis.md>): Clinical Reasoning; MedR-Bench Diagnosis; Accuracy: 93.31.

- [Clinical Reasoning on MedR-Bench Treatment](<https://www.sota2.com/research/sota/clinical-reasoning-on-medr-bench-treatment.md>): Clinical Reasoning; MedR-Bench Treatment; Accuracy: 89.19.

- [Diagnosis on MedR-Bench](<https://www.sota2.com/research/sota/diagnosis-on-medr-bench.md>): Diagnosis; MedR-Bench; R: 77.

- [Treatment on MedR-Bench](<https://www.sota2.com/research/sota/treatment-on-medr-bench.md>): Treatment; MedR-Bench; R Score: 77.3.

- [Medical Question Answering on MedAgentsBench](<https://www.sota2.com/research/sota/medical-question-answering-on-medagentsbench.md>): Medical Question Answering; MedAgentsBench; MedBullets Score: 43.82.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2511.06449>)

[PDF](<https://arxiv.org/pdf/2511.06449.pdf>)