LifeBench: A Benchmark for Long-Horizon Multi-Source Memory

TL;DR

LifeBench is a new benchmark designed to evaluate AI agents' ability to perform long-horizon, multi-source memory reasoning that includes both declarative and non-declarative types, using diverse, real-world inspired data.

Contribution

This paper introduces LifeBench, a comprehensive benchmark that incorporates long-term, multi-source memory tasks with real-world data, addressing limitations of existing memory benchmarks.

Findings

State-of-the-art systems achieve only 55.2% accuracy on LifeBench.

LifeBench effectively challenges memory systems with long-horizon, multi-source reasoning tasks.

The dataset emphasizes data quality, diversity, and scalability, inspired by cognitive science principles.

Abstract

Long-term memory is fundamental for personalized agents capable of accumulating knowledge, reasoning over user experiences, and adapting across time. However, existing memory benchmarks primarily target declarative memory, specifically semantic and episodic types, where all information is explicitly presented in dialogues. In contrast, real-world actions are also governed by non-declarative memory, including habitual and procedural types, and need to be inferred from diverse digital traces. To bridge this gap, we introduce Lifebench, which features densely connected, long-horizon event simulation. It pushes AI agents beyond simple recall, requiring the integration of declarative and non-declarative memory reasoning across diverse and temporally extended contexts. Building such a benchmark presents two key challenges: ensuring data quality and scalability. We maintain data quality by employing real-world priors, including anonymized social surveys, map APIs, and holiday-integrated calendars, thus enforcing fidelity, diversity and behavioral rationality within the dataset. Towards scalability, we draw inspiration from cognitive science and structure events according to their partonomic hierarchy; enabling efficient parallel generation while maintaining global coherence. Performance results show that top-tier, state-of-the-art memory systems reach just 55.2\% accuracy, highlighting the inherent difficulty of long-horizon retrieval and multi-source integration within our proposed benchmark. The dataset and data synthesis code are available at https://github.com/1754955896/LifeBench.