DeepSearchQA: Bridging the Comprehensiveness Gap for Deep Research Agents

TL;DR

DeepSearchQA introduces a challenging benchmark with 900 multi-step, open-web information-seeking tasks across 17 fields, exposing current agent limitations in complex search, reasoning, and answer precision.

Contribution

The paper presents a novel benchmark dataset designed to evaluate deep research agents on complex, multi-step information retrieval and reasoning tasks, highlighting current performance gaps.

Findings

State-of-the-art agents struggle with balancing recall and precision.

Agents often prematurely stop or over-generate answers, revealing weaknesses in planning and confidence estimation.

DeepSearchQA serves as a diagnostic tool for improving deep research agent capabilities.

Abstract

We introduce DeepSearchQA, a 900-prompt benchmark for evaluating agents on difficult multi-step information-seeking tasks across 17 different fields. Unlike traditional benchmarks that target single answer retrieval or broad-spectrum factuality, DeepSearchQA features a dataset of challenging, handcrafted tasks designed to evaluate an agent's ability to execute complex search plans to generate exhaustive answer lists. This shift in design explicitly tests three critical, yet under-evaluated capabilities: 1) systematic collation of fragmented information from disparate sources, 2) de-duplication and entity resolution to ensure precision, and 3) the ability to reason about stopping criteria within an open-ended search space. Each task is structured as a causal chain, where discovering information for one step is dependent on the successful completion of the previous one, stressing long-horizon planning and context retention. All tasks are grounded in the open web with objectively verifiable answer sets. Our comprehensive evaluation of state-of-the-art agent architectures reveals significant performance limitations: even the most advanced models struggle to balance high recall with precision. We observe distinct failure modes ranging from premature stopping (under-retrieval) to hedging behaviors, where agents cast an overly wide net of low-confidence answers to artificially boost recall. These findings highlight critical headroom in current agent designs and position DeepSearchQA as an essential diagnostic tool for driving future research toward more robust, deep-research capabilities.