WorldTravel: A Realistic Multimodal Travel-Planning Benchmark with Tightly Coupled Constraints

Real-world autonomous planning requires coordinating tightly coupled constraints where a single decision dictates the feasibility of all subsequent actions. However, existing benchmarks predominantly feature loosely coupled constraints solvable through local greedy decisions and rely on idealized data, failing to capture the complexity of extracting parameters from dynamic web environments. We introduce \textbf{WorldTravel}, a benchmark comprising 150 real-world travel scenarios across 5 cities that demand navigating an average of 15+ interdependent temporal and logical constraints. To evaluate agents in realistic deployments, we develop \textbf{WorldTravel-Webscape}, a multi-modal environment featuring over 2,000 rendered webpages where agents must perceive constraint parameters directly from visual layouts to inform their planning. Our evaluation of 10 frontier models reveals a significant performance collapse: even the state-of-the-art GPT-5.2 achieves only 32.67\% feasibility in text-only settings, which plummets to 19.33\% in multi-modal environments. We identify a critical Perception-Action Gap and a Planning Horizon threshold at approximately 10 constraints where model reasoning consistently fails, suggesting that perception and reasoning remain independent bottlenecks. These findings underscore the need for next-generation agents that unify high-fidelity visual perception with long-horizon reasoning to handle brittle real-world logistics.

• Publication year

  2026

• Venue

  Unknown venue

• Publication date

  2026-02-09

• Fields of study

  Computer Science

• Identifiers
  arXiv 2602.08367
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models

  2025cited by this paper
• Simulating Classroom Education with LLM-Empowered Agents

  2024cited by this paper
• ChinaTravel: An Open-Ended Benchmark for Language Agents in Chinese Travel Planning

  2024cited by this paper
• TravelAgent: An AI Assistant for Personalized Travel Planning

  2024cited by this paper
• TravelPlanner: A Benchmark for Real-World Planning with Language Agents

  2024cited by this paper
• ItiNera: Integrating Spatial Optimization with Large Language Models for Open-domain Urban Itinerary Planning

  2024cited by this paper
• Large Language Models Can Solve Real-World Planning Rigorously with Formal Verification Tools

  2024cited by this paper
• Hello Again! LLM-powered Personalized Agent for Long-term Dialogue

  2024cited by this paper
• A survey on large language model based autonomous agents

  2023cited by this paper
• A survey on personalized itinerary recommendation: From optimisation to deep learning

  2023cited by this paper
• Graph of Thoughts: Solving Elaborate Problems with Large Language Models

  2023cited by this paper
• Toolformer: Language Models Can Teach Themselves to Use Tools

  2023cited by this paper
• HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face

  2023cited by this paper
• Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models

  2023cited by this paper
• MemoryBank: Enhancing Large Language Models with Long-Term Memory

  2023cited by this paper
• Tree of Thoughts: Deliberate Problem Solving with Large Language Models

  2023cited by this paper
• HuatuoGPT, towards Taming Language Model to Be a Doctor

  2023cited by this paper
• ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs

  2023cited by this paper
• ReAct: Synergizing Reasoning and Acting in Language Models

  2022cited by this paper
• Chain of Thought Prompting Elicits Reasoning in Large Language Models

  2022cited by this paper
• Gemini

  2009cited by this paper
• Corresponding author.

  2004cited by this paper

• No citing papers are available for this paper.