# RExBench: Can coding agents autonomously implement AI research extensions?

arXiv: https://arxiv.org/abs/2506.22598 · HTML: https://mlsanity.com/paper/2506.22598/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- Placed in Web, search, and research agents, one of 16 directions in LLM agents and tool use.
- Builds on 36 of its 59 references that are in the map.
- First posted on arXiv 27 Jun 2025.

## Placed in

- LLM agents and tool use / [Web, search, and research agents](https://mlsanity.com/agents/web-search-and-research-agents/)

## Builds on (most cited first, 30 of 36)

- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [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)
- [Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task](https://mlsanity.com/paper/2210.13382/) (2022)
- [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)
- [TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks](https://mlsanity.com/paper/2412.14161/) (2024)
- [Emergent Linear Representations in World Models of Self-Supervised Sequence Models](https://mlsanity.com/paper/2309.00941/) (2023)
- [COGS: A Compositional Generalization Challenge Based on Semantic Interpretation](https://mlsanity.com/paper/2010.05465/) (2020)
- [Reasoning or Reciting? Exploring the Capabilities and Limitations of Language Models Through Counterfactual Tasks](https://mlsanity.com/paper/2307.02477/) (2023)
- [SciCode: A Research Coding Benchmark Curated by Scientists](https://mlsanity.com/paper/2407.13168/) (2024)
- [LAB-Bench: Measuring Capabilities of Language Models for Biology Research](https://mlsanity.com/paper/2407.10362/) (2024)
- [SWE-rebench: An Automated Pipeline for Task Collection and Decontaminated Evaluation of Software Engineering Agents](https://mlsanity.com/paper/2505.20411/) (2025)
- [Language agents achieve superhuman synthesis of scientific knowledge](https://mlsanity.com/paper/2409.13740/) (2024)
- [DSBench: How Far Are Data Science Agents from Becoming Data Science Experts?](https://mlsanity.com/paper/2409.07703/) (2024)
- [MLGym: A New Framework and Benchmark for Advancing AI Research Agents](https://mlsanity.com/paper/2502.14499/) (2025)
- [CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark](https://mlsanity.com/paper/2409.11363/) (2024)
- [The Devil is in the Detail: Simple Tricks Improve Systematic Generalization of Transformers](https://mlsanity.com/paper/2108.12284/) (2021)
- [CheckEval: A reliable LLM-as-a-Judge framework for evaluating text generation using checklists](https://mlsanity.com/paper/2403.18771/) (2024)
- [LLM4SR: A Survey on Large Language Models for Scientific Research](https://mlsanity.com/paper/2501.04306/) (2025)
- [BLADE: Benchmarking Language Model Agents for Data-Driven Science](https://mlsanity.com/paper/2408.09667/) (2024)
- [Paper2Code: Automating Code Generation from Scientific Papers in Machine Learning](https://mlsanity.com/paper/2504.17192/) (2025)
- [LLMs Assist NLP Researchers: Critique Paper (Meta-)Reviewing](https://mlsanity.com/paper/2406.16253/) (2024)
- [DataSciBench: An LLM Agent Benchmark for Data Science](https://mlsanity.com/paper/2502.13897/) (2025)
- [CodeScientist: End-to-End Semi-Automated Scientific Discovery with Code-based Experimentation](https://mlsanity.com/paper/2503.22708/) (2025)
- [SUPER: Evaluating Agents on Setting Up and Executing Tasks from Research Repositories](https://mlsanity.com/paper/2409.07440/) (2024)
- [SciReplicate-Bench: Benchmarking LLMs in Agent-driven Algorithmic Reproduction from Research Papers](https://mlsanity.com/paper/2504.00255/) (2025)
- [VariErr NLI: Separating Annotation Error from Human Label Variation](https://mlsanity.com/paper/2403.01931/) (2024)
- [ResearchCodeBench: Benchmarking LLMs on Implementing Novel Machine Learning Research Code](https://mlsanity.com/paper/2506.02314/) (2025)
- [Curie: Toward Rigorous and Automated Scientific Experimentation with AI Agents](https://mlsanity.com/paper/2502.16069/) (2025)

## Built on by (most cited first, 18 of 18)

- [RECODE-H: A Benchmark for Research Code Development with Interactive Human Feedback](https://mlsanity.com/paper/2510.06186/) (2025)
- [Remote Labor Index: Measuring AI Automation of Remote Work](https://mlsanity.com/paper/2510.26787/) (2025)
- [Can AI agents conduct open-ended AI research? Early evidence from two case studies](https://mlsanity.com/paper/2607.27191/) (2026)
- [SWE-bench Science: Can Coding Agents Resolve Engineering Tasks in Science?](https://mlsanity.com/paper/2608.19799/) (2026)
- [ResearchEnvBench: Benchmarking Agents on Environment Synthesis for Research Code Execution](https://mlsanity.com/paper/2603.06739/) (2026)
- [LoopsBench: From Harness Engineering to Loop Engineering in Coding Agent Evaluation](https://mlsanity.com/paper/2608.00267/) (2026)
- [Empowering Real-World: A Survey on the Technology, Practice, and Evaluation of LLM-driven Industry Agents](https://mlsanity.com/paper/2510.17491/) (2025)
- [VERITAS: Towards a General-Purpose Replication Tool for Scientific Research](https://mlsanity.com/paper/2607.02931/) (2026)
- [What Makes AI Research Replicable? Executable Knowledge Graphs as Scientific Knowledge Representations](https://mlsanity.com/paper/2510.17795/) (2025)
- [REVERE: Reflective Evolving Research Engineer](https://mlsanity.com/paper/2603.20667/) (2026)
- [AgentFold: Closed-Loop Agentic Search for Protein Folding Model Design](https://mlsanity.com/paper/2608.26747/) (2026)
- [CodeDistiller: Automatically Generating Code Libraries for Scientific Coding Agents](https://mlsanity.com/paper/2512.01089/) (2025)
- [AblateCell: A Reproduce-then-Ablate Agent for Virtual Cell Repositories](https://mlsanity.com/paper/2604.19606/) (2026)
- [When AI Designs AI: Innovation or Imitation?](https://mlsanity.com/paper/2608.17471/) (2026)
- [What Do CAE Simulation Agents Really Need Beyond a Generic Harness?](https://mlsanity.com/paper/2609.03718/) (2026)
- [WhatWorkedBench: Benchmarking Experimental Understanding in AI Agents](https://mlsanity.com/paper/2609.27490/) (2026)
- [RECLAIM: Can Agents Reproduce the Claims of Machine Learning Papers?](https://mlsanity.com/paper/2609.28850/) (2026)
- [Recursive Improvement of a Differentiable Scientific Software Ecosystem](https://mlsanity.com/paper/2610.04561/) (2026)

## Abstract (arXiv)

Agents based on Large Language Models (LLMs) have shown promise for performing sophisticated software engineering tasks autonomously. In addition, there has been progress towards developing agents that can perform parts of the research pipeline in machine learning and the natural sciences. We argue that research extension and its implementation is a critical capability for such systems, and introduce RExBench to support the evaluation of this capability. RExBench is a benchmark consisting of realistic extensions of 12 research papers that aim to investigate novel research hypotheses. Each task is set up as an extension to an existing research paper and codebase, accompanied by domain expert-written instructions. RExBench is robust to data contamination and supports an automatic evaluation infrastructure that executes agent outputs to determine whether the success criteria are met. We use this benchmark to evaluate 12 LLM agents implemented using two different frameworks, aider and OpenHands. We find that all agents fail to autonomously implement the majority of the extensions, with the best agent achieving around a 33% success rate. Although the success rate improves with additional human-written hints, the best performance under this setting remains below 44%. This indicates that current agents are still short of being able to handle realistic research extension tasks without substantial human guidance.