Multimodal retrieval is becoming a crucial component of modern AI applications, yet its evaluation lags behind the demands of more realistic and challenging scenarios. Existing benchmarks primarily probe surface-level semantic correspondence (e.g., object–text matching) while failing to assess the deeper reasoning required to capture complex relationships between visual and textual information. To address this gap, we introduce MR²-Bench, a reasoning-intensive benchmark for multimodal retrieval. MR²-Bench presents the following critical values: 1) all tasks are reasoning-driven, going beyond shallow matching to effectively assess models’ capacity for logical, spatial, and causal inference; 2) it features diverse multimodal data, such as natural images, diagrams, and visual puzzles, enabling comprehensive evaluation across content types; 3) it supports complex queries and documents containing multiple images and covers diverse retrieval scenarios, more accurately reflecting real-world applications. Our benchmark contains 1,309 curated queries, derived either from manual collection and annotation or from selective consolidation of public datasets. Despite achieving strong results on existing benchmarks, current state-of-the-art models still struggle on MR²-Bench: for example, the leading Seed1.6-Embedding model attains a Recall@1 of 77.78 on MMEB, but only 9.91 on MR²-Bench. This substantial performance gap highlights both the increased challenge posed by our benchmark and the pressing need for further advances in reasoning-intensive multimodal retrieval.

Despite the rapidly growing demand for multimodal retrieval, progress in this field remains severely constrained by a lack of training data. In this paper, we introduce MegaPairs, a novel data synthesis method that leverages vision language models (VLMs) and open-domain images, together with a massive synthetic dataset generated from this method. Our empirical analysis shows that MegaPairs generates high-quality data, enabling the multimodal retriever to significantly outperform the baseline model trained on 70× more data from existing datasets. Moreover, since MegaPairs solely relies on general image corpora and open-source VLMs, it can be easily scaled up, enabling continuous improvements in retrieval performance. In this stage, we produced more than 26 million training instances and trained several models of varying sizes using this data. These new models achieve state-of-the-art zero-shot performance across 4 popular composed image retrieval (CIR) benchmarks and the highest overall performance on the 36 datasets provided by MMEB. They also demonstrate notable performance improvements with additional downstream fine-tuning. Our code, synthesized dataset, and pre-trained models are publicly available at https://github.com/VectorSpaceLab/MegaPairs.

Visual Information Extraction (VIE), as a crucial task of Document Intelligence, involves two primary sub-tasks: Semantic Entity Recognition (SER) and Relation Extraction (RE). However, VIE faces two significant challenges. Firstly, most existing models inadequately utilize spatial information of entities, often failing to predict connections or incorrectly linking spatially distant entities. Secondly, the improper input order of tokens challenges in extracting complete entity pairs from documents with multi-line entities when text is extracted via PDF parser or OCR. To address these challenges, we propose LayoutPointer, a Spatial-Context Adaptive Pointer Network. LayoutPointer explicitly enhances spatial-context relationships by incorporating 2D relative position information and adaptive spatial constraints within self-attention. Furthermore, we recast the RE task as a specialized cycle detection problem, employing a unique tail-to-head pointer to restore the semantic order among multi-line entities. To better evaluate the effectiveness of our proposed method, we reconstruct a multi-line dataset named MLFUD, which more accurately reflects real-world scenarios. Fine-tuning experimental results on FUNSD, XFUND, and MLFUD datasets demonstrate that LayoutPointer significantly outperforms existing state-of-the-art methods in F1 scores for RE tasks (e.g., 5.71% improvement on XFUND using LayoutPointer_BASE-X over LayoutLMv3).

Multi-modal retrieval becomes increasingly popular in practice. However, the existing retrievers are mostly text-oriented, which lack the capability to process visual information. Despite the presence of vision-language models like CLIP, the current methods are severely limited in representing the text-only and image-only data. In this work, we present a new embedding model VISTA for universal multi-modal retrieval. Our work brings forth threefold technical contributions. Firstly, we introduce a flexible architecture which extends a powerful text encoder with the image understanding capability by introducing visual token embeddings. Secondly, we develop two data generation strategies, which bring high-quality composed image-text to facilitate the training of the embedding model. Thirdly, we introduce a multi-stage training algorithm, which first aligns the visual token embedding with the text encoder using massive weakly labeled data, and then develops multi-modal representation capability using the generated composed image-text data. In our experiments, VISTA achieves superior performances across a variety of multi-modal retrieval tasks in both zero-shot and supervised settings. Our model, data, and source code are available at https://github.com/FlagOpen/FlagEmbedding.