aiqbench analysis
This paper presents EvE, a decentralized framework that co-evolves coding agents and their guidance states to overcome static performance limits in algorithmic discovery. By dynamically adapting agent behaviors based on marginal gains rather than relying on frozen models, the system successfully resolved bottlenecks in In-Context Operator Networks through autonomous mechanism discovery. The findings underscore the critical role of stage-dependent adaptation in navigating complex search landscapes for advanced AI agents.
Evolving Ensemble of Agents
arXiv:2605.09018v4 Announce Type: replace-cross Abstract: We introduce the Evolving Ensemble of Agents (EvE), a decentralized framework that organizes existing, highly capable coding agents into a live, co-evolving system for algorithmic discovery. Rather than reinventing the wheel within the ``LLMs as optimizers'' paradigm, EvE fixes the base agent substrate and focuses entirely on evolving the cumulative guidance and skills that dictate agent behaviors. By maintaining two co-evolving populations, namely functional code solvers and agent guidance states, the system evaluates agents through a synchronous race, updating their empirical Elo ratings based on the marginal gains they contribute to the current solver state. When applied to a research bottleneck in In-Context Operator Networks (ICON), EvE autonomously discovered a robust rescale-then-interpolate mechanism that enables reliable example-count generalization. Crucially, controlled ablations reveal the absolute necessity of stage-dependent agent adaptation to navigate the shifting search landscapes of complex codebases. Compared to variants driven by a fixed initial agent or even a frozen ``best-evolved'' agent, EvE uniquely avoids phase mismatch, demonstrating that organizing agents into a self-revising ensemble is the fundamental driver for breaking through static performance ceilings.
Abstract:We introduce the Evolving Ensemble of Agents (EvE), a decentralized framework that organizes existing, highly capable coding agents into a live, co-evolving system for algorithmic discovery. Rather than reinventing the wheel within the ``LLMs as optimizers'' paradigm, EvE fixes the base agent substrate and focuses entirely on evolving the cumulative guidance and skills that dictate agent behaviors. By maintaining two co-evolving populations, namely functional code solvers and agent guidance states, the system evaluates agents through a synchronous race, updating their empirical Elo ratings based on the marginal gains they contribute to the current solver state. When applied to a research bottleneck in In-Context Operator Networks (ICON), EvE autonomously discovered a robust rescale-then-interpolate mechanism that enables reliable example-count generalization. Crucially, controlled ablations reveal the absolute necessity of stage-dependent agent adaptation to navigate the shifting search landscapes of complex codebases. Compared to variants driven by a fixed initial agent or even a frozen ``best-evolved'' agent, EvE uniquely avoids phase mismatch, demonstrating that organizing agents into a self-revising ensemble is the fundamental driver for breaking through static performance ceilings.
Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2605.09018 [cs.NE]
(or arXiv:2605.09018v4 [cs.NE] for this version)
https://doi.org/10.48550/arXiv.2605.09018 Focus to learn more arXiv-issued DOI via DataCite
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.
Full text under the publisher's open licence · 276 words
Related coverage
TestGRAD: Evolving Test Suites via Failure Pattern Momentum for SWE-Agent Ensemble
arXiv · Oct 8, 2026
G\"odel Forest: Balancing Search Depth and Breadth for Data-Centric Recursive Self-Improvement
arXiv · Oct 1, 2026
EpiEvolve: Self-Evolving Agents for Streaming Pandemic Forecasting under Regime Shifts
arXiv · Jun 6, 2026
Co-Evolving Acting and Forecasting for Language Agents
arXiv · Oct 15, 2026