How Far Are We from Genuinely Useful Deep Research Agents?

TL;DR

This paper introduces a new benchmark and failure taxonomy for Deep Research Agents, revealing current limitations in evidence integration and reasoning, and aiming to guide future improvements in automated research report generation.

Contribution

It presents FINDER, a standardized benchmark with structured research tasks, and DEFT, a detailed failure taxonomy for evaluating deep research agents.

Findings

Current DRAs excel at task understanding but struggle with evidence verification.

Most failures occur in reasoning, retrieval, and generation processes.

The benchmark enables more objective evaluation of DRA performance.

Abstract

Deep Research Agents (DRAs) aim to automatically produce analyst-level reports through iterative information retrieval and synthesis. However, most existing DRAs were validated on question-answering benchmarks, while research on generating comprehensive reports remains overlooked. Worse, current benchmarks for report synthesis suffer from task complexity and subjective metrics -- this fails to reflect user demands and limits the practical utility of generated reports. To address these gaps, we present Fine-grained DEepResearch bench (FINDER), an enhanced benchmark consisting of 100 human-curated research tasks with 419 structured checklist items that standardize report structure, analytical depth, and factual grounding. Based on approximately 1,000 reports produced by mainstream DRAs, we further propose Deep rEsearch Failure Taxonomy (DEFT), the first failure taxonomy for deep research agents. DEFT contains 14 fine-grained failure modes across reasoning, retrieval, and generation, and is built upon grounded theory with human-LLM co-annotating and inter-annotator reliability validation. Our experimental findings reveal that current DRAs struggle not with task comprehension but with evidence integration, verification, and reasoning-resilient planning.