Agentic coding without the cloud: evaluating open-weight large language models on longitudinal data preparation tasks

A new open-source framework from University College London (UCL) evaluates open-weight large language models on longitudinal data preparation tasks without sending data to the cloud. Benchmarking models up to 31-35B parameters achieved up to 87.9% average task completion on 20 tasks creating 102 variables, showing promise for AI-assisted data preparation in governance-restricted research settings.

arXiv:2607.21482v2 Announce Type: replace-cross Abstract: Large language models LLMs and agents are now widely used tools in code development, with data typically sent to third-party cloud-based models. Their adoption in research using personal data is constrained by governance requirements that typically prohibit data transmission to external services. Locally deployable open-weight models offer an alternative since sensitive data never leave the local environment. We introduce an open-source framework for evaluating the efficacy of AI agents powered by open-weight LLMs on one of the most persistent bottlenecks in research on longitudinal population studies: data preparation. The framework comprises: a curated ground-truth dataset cleaning scripts preparing six sweeps of data from a British cohort study , task definitions encompassing tasks such as category harmonization and multi-wave merging, and automated routines for evaluating the LLM-produced R code and outputted data. We benchmark LLMs across the consumer grade deployment spectrum to assess their efficacy in 20 data preparation tasks creation of 102 variables . Current state-of-the-art, 31-35B parameter models almost saturated our benchmark 'average task completion' up to 87.9% . The performance of open-weight LLMs running on consumer-grade hardware shows promise of a viable path toward AI-assisted data preparation in governance-restricted research settings. Our framework is publicly available at: https://github.com/UCL-ARC/RRBench.