What Limits Virtual Agent Application? OmniBench: A Scalable Multi-Dimensional Benchmark for Essential Virtual Agent Capabilities

TL;DR

OmniBench introduces a scalable, multidimensional benchmark with automated task synthesis for evaluating and advancing virtual agent capabilities across diverse scenarios.

Contribution

We developed OmniBench and OmniEval, enabling controllable, multidimensional evaluation of virtual agents with synthesized tasks and comprehensive metrics.

Findings

High human acceptance rate of 91% for synthesized tasks

Graph-structured data improves training efficiency

Performance variability across models and capabilities

Abstract

As multimodal large language models (MLLMs) advance, MLLM-based virtual agents have demonstrated remarkable performance. However, existing benchmarks face significant limitations, including uncontrollable task complexity, extensive manual annotation with limited scenarios, and a lack of multidimensional evaluation. In response to these challenges, we introduce OmniBench, a self-generating, cross-platform, graph-based benchmark with an automated pipeline for synthesizing tasks of controllable complexity through subtask composition. To evaluate the diverse capabilities of virtual agents on the graph, we further present OmniEval, a multidimensional evaluation framework that includes subtask-level evaluation, graph-based metrics, and comprehensive tests across 10 capabilities. Our synthesized dataset contains 36k graph-structured tasks across 20 scenarios, achieving a 91\% human acceptance rate. Training on our graph-structured data shows that it can more efficiently guide agents compared to manually annotated data. We conduct multidimensional evaluations for various open-source and closed-source models, revealing their performance across various capabilities and paving the way for future advancements. Our project is available at https://omni-bench.github.io/.