VLNVerse: A Benchmark for Vision-Language Navigation with Versatile, Embodied, Realistic Simulation and Evaluation

TL;DR

VLNVerse introduces a large-scale, versatile benchmark for vision-language navigation that unifies tasks, supports realistic embodied simulation, and enables comprehensive evaluation and development of general-purpose embodied agents.

Contribution

It presents VLNVerse, a scalable, unified benchmark with realistic physics simulation, and proposes a new multi-task model to address all tasks within this framework.

Findings

Comprehensive evaluation of existing VLN methods.

Demonstration of the benchmark's scalability and versatility.

Development of a unified multi-task navigation model.

Abstract

Despite remarkable progress in Vision-Language Navigation (VLN), existing benchmarks remain confined to fixed, small-scale datasets with naive physical simulation. These shortcomings limit the insight that the benchmarks provide into sim-to-real generalization, and create a significant research gap. Furthermore, task fragmentation prevents unified/shared progress in the area, while limited data scales fail to meet the demands of modern LLM-based pretraining. To overcome these limitations, we introduce VLNVerse: a new large-scale, extensible benchmark designed for Versatile, Embodied, Realistic Simulation, and Evaluation. VLNVerse redefines VLN as a scalable, full-stack embodied AI problem. Its Versatile nature unifies previously fragmented tasks into a single framework and provides an extensible toolkit for researchers. Its Embodied design moves beyond intangible and teleporting "ghost" agents that support full-kinematics in a Realistic Simulation powered by a robust physics engine. We leverage the scale and diversity of VLNVerse to conduct a comprehensive Evaluation of existing methods, from classic models to MLLM-based agents. We also propose a novel unified multi-task model capable of addressing all tasks within the benchmark. VLNVerse aims to narrow the gap between simulated navigation and real-world generalization, providing the community with a vital tool to boost research towards scalable, general-purpose embodied locomotion agents.