@inproceedings{kim-etal-2026-omhbench,
    title = "{OMHB}ench: Benchmarking Balanced and Grounded Omni-Modal Multi-Hop Reasoning",
    author = "Kim, Seunghee  and
      Bang, Ingyu  and
      Jang, Seokgyu  and
      Kim, Changhyeon  and
      Bae, Sanghwan  and
      Choi, Jihun  and
      Xuan, Richeng  and
      Kim, Taeuk",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.911/",
    pages = "18311--18334",
    ISBN = "979-8-89176-395-1",
    abstract = "Multimodal Large Language Models (MLLMs) have increasingly supported omni-modal processing across text, vision, and speech. However, existing evaluation frameworks for such models suffer from critical limitations, including modality shortcuts and biased reasoning paths. To address these challenges, we propose OMHBench, a novel benchmark designed to rigorously evaluate omni-modal multi-hop reasoning. It consists of 6,144 questions with balanced reasoning paths that are jointly grounded across all three modalities. Extensive evaluation of 13 state-of-the-art models reveals that (1) a large performance gap exists between proprietary and open-source MLLMs and (2) even proprietary models exhibit high sensitivity to reasoning path variations, resulting in asymmetric omni-modal grounding. Notably, models struggle when processing the speech modality, underscoring the need for balanced, multi-hop evaluation of omni-modal intelligence."
}