What We are Missing in Multimodal LLM Evaluation?
For researchers developing and evaluating multimodal LLMs, the paper highlights critical missing dimensions in assessment that limit understanding of model capabilities.
The paper identifies gaps in current multimodal LLM evaluation benchmarks, such as temporal-spatial coherence and multimodal consistency, arguing that existing tests fail to measure true cross-modal integration.
Multimodal large language models (MLLMs) can process diverse inputs, e.g., text, images, audio, and video, and generate textual responses. While their capabilities have advanced rapidly, evaluation of such models has not kept pace. Most existing evaluation benchmarks are limited to isolated tasks and reveal little about whether a model integrates information across modalities. We examine current means for evaluating MLLMs and review the existing benchmark taxonomy to identify gaps, including temporal-spatial coherence, physical world understanding, multimodal consistency, and selective attention. Addressing these gaps is essential for measuring real progress in multimodal intelligence and exposing capability boundaries.