analiza aiqbench
This study demonstrates that system architecture, rather than raw model intelligence, is the primary determinant of performance in multi-agent LLM frameworks. By isolating architectural variables across nine frameworks, the research reveals how specific design choices like orchestration overhead and communication topology drastically impact latency and coordination success rates. The findings provide a critical taxonomy and evidence-based principles for engineers optimizing complex agent systems beyond simple model selection.
Architectural Design, Not Only Model Intelligence, Governs Multi-Agent LLM Performance
arXiv:2602.03128v2 Announce Type: replace Abstract: Multi-agent LLM frameworks are data-intensive systems that govern how agents orchestrate tasks, manage state, and coordinate decisions. These architectural choices control execution overhead, memory behavior, planning effectiveness, and coordination scalability. Their impact on system performance remains poorly understood. Existing benchmarks evaluate individual agent capabilities in isolation and lack standardized framework-level comparison. We make four contributions. We introduce an architectural taxonomy that decomposes multi-agent LLM frameworks along five dimensions: orchestration, memory, planning interfaces, specialization, and communication topology. We develop MAFBench, a unified evaluation suite that integrates existing benchmarks within a standardized execution pipeline. We conduct a controlled empirical study across nine frameworks, fixing the underlying LLM and varying only architectural design choices. We distill the results into six evidence-based design principles. Architectural design, not only model intelligence, governs performance. Orchestration alone increases latency by over 60x, and a minimal implementation of the same paradigm isolates that cost as implementation rather than paradigm. Schema-constrained planning interfaces reduce accuracy by up to 32 points through formatting failures, not reasoning errors. Communication topology drops coordination success from above 90% to below 30% under mismatched structure. Memory architecture controls recall and scalability independent of context window size, and no evaluated framework natively supports controlled knowledge revision.
Abstract:Multi-agent LLM frameworks are data-intensive systems that govern how agents orchestrate tasks, manage state, and coordinate decisions. These architectural choices control execution overhead, memory behavior, planning effectiveness, and coordination scalability. Their impact on system performance remains poorly understood. Existing benchmarks evaluate individual agent capabilities in isolation and lack standardized framework-level comparison. We make four contributions. We introduce an architectural taxonomy that decomposes multi-agent LLM frameworks along five dimensions: orchestration, memory, planning interfaces, specialization, and communication topology. We develop MAFBench, a unified evaluation suite that integrates existing benchmarks within a standardized execution pipeline. We conduct a controlled empirical study across nine frameworks, fixing the underlying LLM and varying only architectural design choices. We distill the results into six evidence-based design principles. Architectural design, not only model intelligence, governs performance. Orchestration alone increases latency by over 60x, and a minimal implementation of the same paradigm isolates that cost as implementation rather than paradigm. Schema-constrained planning interfaces reduce accuracy by up to 32 points through formatting failures, not reasoning errors. Communication topology drops coordination success from above 90% to below 30% under mismatched structure. Memory architecture controls recall and scalability independent of context window size, and no evaluated framework natively supports controlled knowledge revision.
Artificial Intelligence (cs.AI)
I.2.11; C.4; H.3.3; I.2.7
arXiv:2602.03128 [cs.AI]
(or arXiv:2602.03128v2 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2602.03128 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.
Text integral sub licența deschisă a publicației · 297 cuvinte
Subiecte conexe
Berkeley and Heiserman as an Unexhausted Architecture for Embodied Machine Intelligence
arXiv · 21 iul. 2026
.webp%3Fssl%3D1&w=320)
Mallory Unifies Threat Intelligence, Exposure Context, and Response Into One Architecture for Security Teams
Cybersecurity News · 4 aug. 2026
AINTMA: Agentic AI Architecture for Autonomous Test Management with Generative Intelligence, Secure Cloud Communication and Adaptive Quality Analytics
arXiv · 24 iul. 2026
Learning-to-Optimize as the Missing Architectural Layer of AI-Native Networks
arXiv · 22 sept. 2026