Multi-Stream LLMs: Unblocking Language Models with Parallel Streams of Thoughts, Inputs and Outputs

Guinan Su, Yanwu Yang, Xueyan Li, Jonas Geiping

Abstract

The continued improvements in language model capability have unlocked their widespread use as drivers of autonomous agents, for example in coding or computer use applications. However, the core of these systems has not changed much since early instruction-tuned models like ChatGPT. Even advanced AI agents function on message exchange formats, successively exchanging messages with users, systems, with itself (i.e. chain-of-thought) and tools in a single stream of computation. This bottleneck to a

Categories

Cite This Resource

@article{llmsec202600166,
  title = {Multi-Stream LLMs: Unblocking Language Models with Parallel Streams of Thoughts, Inputs and Outputs},
  author = {Guinan Su and Yanwu Yang and Xueyan Li and Jonas Geiping},
  year = {2026},
  url = {https://arxiv.org/abs/2605.12460},
}