OmniBench: Towards The Future of Universal Omni-Language Models
This work addresses the problem of evaluating and improving multimodal AI models for researchers and developers, though it is incremental as it builds on existing MLLM frameworks by focusing on tri-modal integration.
The paper tackles the lack of comprehensive benchmarks for evaluating multimodal large language models (MLLMs) that process visual, acoustic, and textual inputs simultaneously, introducing OmniBench and finding that open-source models perform poorly (below 50% accuracy) in tri-modal reasoning. It also curates OmniInstruct, an 84.5K-sample dataset, to address these limitations and advocates for improved training strategies.
Recent advancements in multimodal large language models (MLLMs) have aimed to integrate and interpret data across diverse modalities. However, the capacity of these models to concurrently process and reason about multiple modalities remains underexplored, partly due to the lack of comprehensive modality-wise benchmarks. We introduce OmniBench, a novel benchmark designed to rigorously evaluate models' ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. We define language models capable of such tri-modal processing as the omni-language models (OLMs). OmniBench is distinguished by high-quality human annotations, ensuring that accurate responses require integrated understanding and reasoning across all three modalities. Our main findings reveal that: i) open-source OLMs exhibit critical limitations in instruction-following and reasoning capabilities within tri-modal contexts; and ii) most baselines models perform poorly (below 50% accuracy) even when provided with alternative textual representations of images or/and audio. These results suggest that the ability to construct a consistent context from text, image, and audio is often overlooked in existing MLLM training paradigms. To address this gap, we curate an instruction tuning dataset of 84.5K training samples, OmniInstruct, for training OLMs to adapt to tri-modal contexts. We advocate for future research to focus on developing more robust tri-modal integration techniques and training strategies to enhance OLMs. Codes, data and live leaderboard could be found at https://m-a-p.ai/OmniBench.