GISA: A Benchmark for General Information-Seeking Assistant

TL;DR

GISA is a new benchmark designed to evaluate general information-seeking assistants using realistic queries, structured answer formats, and process-level supervision, revealing significant performance gaps in current models.

Contribution

The paper introduces GISA, a comprehensive benchmark with human-crafted queries, diverse answer formats, and process trajectories to better evaluate and develop information-seeking LLMs.

Findings

Current models achieve only 19.30% exact match on GISA.

Performance drops on tasks requiring complex reasoning and broad information gathering.

GISA exposes substantial room for improvement in LLM-based search agents.

Abstract

The advancement of large language models (LLMs) has significantly accelerated the development of search agents capable of autonomously gathering information through multi-turn web interactions. Various benchmarks have been proposed to evaluate such agents. However, existing benchmarks often construct queries backward from answers, producing unnatural tasks misaligned with real-world needs. Moreover, these benchmarks tend to focus on either locating specific information or aggregating information from multiple sources, while relying on static answer sets prone to data contamination. To bridge these gaps, we introduce GISA, a benchmark for General Information-Seeking Assistants comprising 373 human-crafted queries that reflect authentic information-seeking scenarios. GISA features four structured answer formats (item, set, list, and table), enabling deterministic evaluation. It integrates both deep reasoning and broad information aggregation within unified tasks, and includes a live subset with periodically updated answers to resist memorization. Notably, GISA provides complete human search trajectories for every query, offering gold-standard references for process-level supervision and imitation learning. Experiments on mainstream LLMs and commercial search products reveal that even the best-performing model achieves only 19.30\% exact match score, with performance notably degrading on tasks requiring complex planning and comprehensive information gathering. These findings highlight substantial room for future improvement.