ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling

Abstract

Large Reasoning Models (LRMs) improve performance by allocating additional inference-time compute to generate extended chain-of-thought reasoning. However, recent studies reveal that sequential test-time scaling often yields diminishing or even negative returns, as longer traces exhibit increased uncertainty, error compounding, and drift from the original problem.

We propose ThinkRetrieve, a test-time scaling framework that augments the reasoning traces of LRMs with dynamically retrieved solved examples at each reasoning step. Given an external corpus of problems paired with step-by-step solutions, ThinkRetrieve retrieves relevant exemplars at each intermediate step and injects them directly into the thinking trace, providing the model with guidance on how to reason rather than merely what facts are relevant.

Experiments across five reasoning models (1.5B–8B parameters) on GSM-8K, MATH-500, AIME 2025, and SciQ demonstrate that ThinkRetrieve consistently improves accuracy over standard test-time scaling, with relative gains of up to 60% on AIME 2025.

Author-written abstract from the arXiv record. See the full paper for the evaluation protocol, baselines, and limitations.

Code and reproducibility

The repository includes the Python library and runnable examples. Start with the tutorial or the four-method comparison script.

The science-agent article is a separate exploratory six-case demonstration with synthetic incidents and private tool routes, selected after exploratory screening. Its results are not a general agent accuracy estimate or a biological safety benchmark.

Cite this paper

Vaibhav Singh, Soumya Suvra Ghosal, Sarvesh Gharat, Soumyabrata Pal, Ramasuri Narayanam, and Dinesh Manocha (2026). ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling. arXiv:2608.10928. https://doi.org/10.48550/arXiv.2608.10928.

Download BibTeXDownload RIS (Zotero / EndNote)

@article{thinkretrieve2026,
  title   = {ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling},
  author  = {Singh, Vaibhav and Ghosal, Soumya Suvra and Gharat, Sarvesh and
             Pal, Soumyabrata and Narayanam, Ramasuri and Manocha, Dinesh},
  journal = {arXiv preprint arXiv:2608.10928},
  year    = {2026},
  doi     = {10.48550/arXiv.2608.10928},
  url     = {https://arxiv.org/abs/2608.10928}
}