# Adversarial Speculative Reasoning Improves Long Horizon Agents

Authors: Hongyin Luo, Wei Fang, Lijie Yang, Owen Stepan, Daniel Suit, Jeff Krapf, Dana Wensberg, Jack O'Brien.

Measures how numerical errors accumulate across speculative verification blocks. Trains a draft model to trigger rejection near decisions in a written reasoning trace, restarting verification where cache entries are computed with higher precision.

First public: 2026-11-03.
ArXiv initial submission: 2026-11-03.

[Original paper](https://arxiv.org/abs/2611.05345) · [PDF](https://arxiv.org/pdf/2611.05345) · [HTML](https://arxiv.org/html/2611.05345v1) · [Reading-list entry](https://latent-implicit-thinking.github.io/papers/2611.05345/)
[Preprint BibTeX](https://latent-implicit-thinking.github.io/papers/2611.05345/citation.bib) · [Preprint CSL-JSON](https://latent-implicit-thinking.github.io/papers/2611.05345/citation.csl.json) · [JSON record](https://latent-implicit-thinking.github.io/papers/2611.05345/record.json)

## Approaches

- [Alternative generation methods](https://latent-implicit-thinking.github.io/topics/alternative-reasoning-schedules/): Change how intermediate steps are generated, revised, or searched—for example, by refining several steps in parallel. Some produce written chains of thought.
- [Analysis and evaluation](https://latent-implicit-thinking.github.io/topics/implicit-reasoning-evaluation/): Measure how well these methods work, investigate what their internal computation does, and identify where they fail. This includes experiments, studies of model internals, and mathematical results.

Reading-list version 0.15.2; literature checked through 2026-11-03.