Current Large Language Model reasoning systems process queries independently, discarding valuable cross-instance signals such as shared reasoning patterns and consistency constraints. We introduce Batch-of-Thought (BoT), a training-free method that processes related queries jointly to enable cross-instance learning. By performing comparative analysis across batches, BoT identifies high-quality reasoning templates, detects errors through consistency checks, and amortizes computational costs. We instantiate BoT within a multi-agent reflection architecture (BoT-R), where a Reflector performs joint evaluation to unlock mutual information gain unavailable in isolated processing. Experiments across three model families and six benchmarks demonstrate that BoT-R consistently improves accuracy and confidence calibration while reducing inference costs by up to 61%. Our theoretical and experimental analysis reveals when and why batch-aware reasoning benefits LLM systems. Our code is available at https://github.com/xuanyang19/BoT
@article{arxiv.2601.02950,
title = {Batch-of-Thought: Cross-Instance Learning for Enhanced LLM Reasoning},
author = {Xuan Yang and Furong Jia and Roy Xie and Xiong Xi and Hengwei Bian and Jian Li and Monica Agrawal},
journal= {arXiv preprint arXiv:2601.02950},
year = {2026}
}