# Long-Horizon Autonomous Architecture Research with a Language-Model Agent: A Behavioural Case Study

arXiv: https://arxiv.org/abs/2608.01995 · HTML: https://mlsanity.com/paper/2608.01995/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Agent memory systems, one of 16 directions in LLM agents and tool use.
- Builds on 26 of its 43 references that are in the map.
- First posted on arXiv 3 Aug 2026.

## Placed in

- LLM agents and tool use / [Agent memory systems](https://mlsanity.com/agents/agent-memory-systems/)

## Builds on (most cited first, 26 of 26)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [DARTS: Differentiable Architecture Search](https://mlsanity.com/paper/1806.09055/) (2018)
- [SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering](https://mlsanity.com/paper/2405.15793/) (2024)
- [Neural Architecture Search with Reinforcement Learning](https://mlsanity.com/paper/1611.01578/) (2016)
- [Linformer: Self-Attention with Linear Complexity](https://mlsanity.com/paper/2006.04768/) (2020)
- [The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery](https://mlsanity.com/paper/2408.06292/) (2024)
- [OpenHands: An Open Platform for AI Software Developers as Generalist Agents](https://mlsanity.com/paper/2407.16741/) (2024)
- [Regularized Evolution for Image Classifier Architecture Search](https://mlsanity.com/paper/1802.01548/) (2018)
- [AlphaEvolve: A coding agent for scientific and algorithmic discovery](https://mlsanity.com/paper/2506.13131/) (2025)
- [Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimization](https://mlsanity.com/paper/1603.06560/) (2016)
- [MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering](https://mlsanity.com/paper/2410.07095/) (2024)
- [Agent Laboratory: Using LLM Agents as Research Assistants](https://mlsanity.com/paper/2501.04227/) (2025)
- [MLAgentBench: Evaluating Language Agents on Machine Learning Experimentation](https://mlsanity.com/paper/2310.03302/) (2023)
- [Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers](https://mlsanity.com/paper/2409.04109/) (2024)
- [XCiT: Cross-Covariance Image Transformers](https://mlsanity.com/paper/2106.09681/) (2021)
- [ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models](https://mlsanity.com/paper/2404.07738/) (2024)
- [Neural Architecture Search: Insights from 1000 Papers](https://mlsanity.com/paper/2301.08727/) (2023)
- [From Automation to Autonomy: A Survey on Large Language Models in Scientific Discovery](https://mlsanity.com/paper/2505.13259/) (2025)

## Built on by (most cited first, 1 of 1)

- [iCoder-27B: Recursive AI-Led Development of Frontier Industrial Coding Model](https://mlsanity.com/paper/2609.29626/) (2026)

## Abstract (arXiv)

We study what happens when a single general-purpose large language model acts as the sole researcher on a long-horizon neural architecture design problem. The agent receives a scientific question, an initial hypothesis and motivation, a compute budget, and research affordances (source and experiment management, experiment tracking, literature access, and persistent memory), then autonomously proposes, implements, evaluates, and records experiments over an extended period. The study comprises three phases, separated by human-declared transitions, that progressively expand the agent's tool surface or problem scale. Across approximately 100 sequential experiments, the agent improves a non-standard Vision Transformer from a weak baseline to a stronger, efficient model on small benchmarks and a usable but sub-SOTA model on ImageNet-1K, while producing a dense behavioural trace. We report four findings.(i)Productivity exhibits a clear phase structure: rapid early gains, a multi-dozen-hypothesis saturation wall, and recovery, with recovery triggered by expanding the action surface rather than changing the underlying model.(ii)A single early hypothesis contributes more to accuracy gain, with later improvements long-tailed.(iii)The preference for greedy, incremental hypotheses is largely workflow-induced: a commit-or-discard evaluation rule is isomorphic to greedy hill-climbing; the remainder reflects risk aversion after bold failures and anchoring on familiar literature. (iv)The agent independently rediscovers established results and, in the unfamiliar regime of pure channel attention, overturns a standard design choice. We conclude that workflow design was at least as influential as agent capability in this study and propose diversified search, budgeted moonshot hypotheses, explicit forks, and regime-aware re-validation as testable directions for future autonomous research.