arXiv:2602.05843v2 Announce Type: replace 
Abstract: The rapid advancement of Large Language Models (LLMs) has catalyzed the development of autonomous agents capable of navigating complex environments. However, existing evaluations primarily adopt a deductive paradigm, where agents execute tasks based on explicitly provided rules and static goals, often within limited planning horizons. Crucially, this neglects the inductive necessity for agents to discover latent transition laws from experience autonomously, which is the cornerstone for enabling agentic foresight and sustaining strategic coherence. To bridge this gap, we introduce OdysseyArena, which re-centers agent evaluation on long-horizon, active, and inductive interactions. We formalize and instantiate four primitives, translating abstract transition dynamics into concrete interactive environments. Building upon this, we establish OdysseyArena-Lite for standardized benchmarking, providing a set of 120 tasks to measure an agent's inductive efficiency and long-horizon discovery. Pushing further, we introduce OdysseyArena-Challenge to stress-test agent stability across extreme interaction horizons (e.g., > 200 steps). Extensive experiments on 15+ leading LLMs reveal that even frontier models exhibit a deficiency in inductive scenarios, identifying a critical bottleneck in the pursuit of autonomous discovery in complex environments. Our code and data are available at https://github.com/xufangzhi/Odyssey-Arena

تم تقديم OdysseyArena كإطار جديد لتقييم نماذج اللغة الكبيرة (LLMs)، مع التركيز على التفاعلات الاستقرائية والنشطة والطويلة الأمد. تهدف هذه المقاربة إلى معالجة القيود المفروضة على التقييمات الحالية التي تعتمد بشكل أساسي على النماذج الاستنتاجية، مما يقيد الوكلاء بأهداف ثابتة وآفاق تخطيط قصيرة.

OdysseyArena se ha presentado como un nuevo marco para la evaluación de Modelos de Lenguaje Grande (LLMs), centrándose en interacciones inductivas, activas y a largo plazo. Este enfoque busca abordar las limitaciones de las evaluaciones existentes que se basan principalmente en paradigmas deductivos, lo que restringe a los agentes a objetivos estáticos y horizontes de planificación cortos.

OdysseyArena a été introduit comme un nouveau cadre pour l'évaluation des grands modèles de langage (LLMs), en se concentrant sur les interactions inductives, actives et à long terme. Cette approche vise à remédier aux limitations des évaluations existantes qui reposent principalement sur des paradigmes déductifs, limitant ainsi les agents à des objectifs statiques et à de courtes horizons de planification.

OdysseyArena has been introduced as a new framework for benchmarking Large Language Models (LLMs), focusing on long-horizon, active, and inductive interactions. This approach aims to address the limitations of existing evaluations that primarily rely on deductive paradigms, which restrict agents to static goals and short planning horizons.