---
title: "StressEval: Failure-Driven Dynamic Benchmarking for Knowledge-Intensive Reasoning in Large Language Models"
authors:
  - "Yongrui Chen"
  - "Yangyang Ma"
  - "Xiaoying Huang"
  - "Shenyu Zhang"
  - "Huajun Chen"
  - "Haofen Wang"
  - "Guilin Qi"
doi: "10.24963/ijcai.2026/422"
url: "https://ijcai.org/proceedings/2026/422"
pdf: "https://ijcai.org/proceedings/2026/0422.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "3792-3800"
track: "Main Track"
keywords:
  - "Knowledge Representation and Reasoning: Learning and reasoning"
  - "Natural Language Processing: Applications"
  - "Natural Language Processing: Question answering"
---

# StressEval: Failure-Driven Dynamic Benchmarking for Knowledge-Intensive Reasoning in Large Language Models

Yongrui Chen, Yangyang Ma, Xiaoying Huang, Shenyu Zhang, Huajun Chen, Haofen Wang, Guilin Qi

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 3792-3800. DOI: [10.24963/ijcai.2026/422](https://doi.org/10.24963/ijcai.2026/422)

## Abstract

Static benchmarks for LLMs are increasingly compromised by contamination and overfitting, especially on knowledge-intensive reasoning tasks. While recent dynamic benchmarks can alleviate staleness, they often increase difficulty at the expense of answerability and controllability. In this paper, we propose StressEval, a failure-driven data synthesis framework that turns observed model failures into dynamic, challenging, and controlable test instances. StressEval consists of three stages: (i) it constructs a semi-structured difficulty card that identifies the failed reasoning step and its root cause; (ii) it applies a dual-perspective instance-synthesis method that targets both knowledge gaps and reasoning breakdowns while preserving the underlying difficulty factors; and (iii) it applies a gating mechanism to retain only grounded, unambiguous instances. Seeding from multiple knowledge-intensive reasoning datasets, we employ StressEval to build Dynamic-OneEval, a focused suite of challenging dynamic benchmark. Across several state-of-the-art LLMs, Dynamic-OneEval yields substantially larger performance drops than the original benchmarks while retaining explicit difficulty factors, enabling more actionable iteration.