# Inference-Time Scaling of Verification: Self-Evolving Deep Research Agents via Test-Time Rubric-Guided Verification Source: [Inference-Time Scaling of Verification: Self-Evolving Deep Research Agents via Test-Time Rubric-Guided Verification]() Collection index: [llms.txt]() Data updated: 2026-05-01T05:00:04.847Z Authors: Yuxuan Wan, Tianqing Fang, Zaitang Li, Yintong Huo, Wenxuan Wang, Haitao Mi, Dong Yu, Michael R. Lyu Published: 2026-01-22T09:47:31.000Z ## Abstract Recent advances in Deep Research Agents (DRAs) are transforming automated knowledge discovery and problem-solving. While the majority of existing efforts focus on enhancing policy capabilities via post-training, we propose an alternative paradigm: self-evolving the agent's ability by iteratively verifying the policy model's outputs, guided by meticulously crafted rubrics. This approach gives rise to the inference-time scaling of verification, wherein an agent self-improves by evaluating its generated answers to produce iterative feedback and refinements. We derive the rubrics based on an automatically constructed DRA Failure Taxonomy, which systematically classifies agent failures into five major categories and thirteen sub-categories. We present DeepVerifier, a rubrics-based outcome reward verifier that leverages the asymmetry of verification and outperforms vanilla agent-as-judge and LLM judge baselines by 12%-48% in meta-evaluation F1 score. To enable practical self-evolution, DeepVerifier integrates as a plug-and-play module during test-time inference. The verifier produces detailed rubric-based feedback, which is fed back to the agent for iterative bootstrapping, refining responses without additional training. This test-time scaling delivers 8%-11% accuracy gains on challenging subsets of GAIA and XBench-DeepSearch when powered by capable closed-source LLMs. Finally, to support open-source advancement, we release DeepVerifier-4K, a curated supervised fine-tuning dataset of 4,646 high-quality agent steps focused on DRA verification. These examples emphasize reflection and self-critique, enabling open models to develop robust verification capabilities. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Deep Research Task on DeepSearch](): Deep Research Task; DeepSearch; Accuracy (%): 47. - [General AI Assistant Reasoning on GAIA (Web)](): General AI Assistant Reasoning; GAIA (Web); Accuracy: 33.33. - [General AI Assistant Reasoning on GAIA Full](): General AI Assistant Reasoning; GAIA (full); Accuracy: 32.21. - [General AI Assistant Reasoning on GAIA (File/Reasoning/Others)](): General AI Assistant Reasoning; GAIA (File/Reasoning/Others); Accuracy: 30.85. - [Deep Research Task on BrowseComp](): Deep Research Task; Browsecomp; Accuracy: 10. ## Original sources and implementation [Original paper]() [PDF]()