# MultiHaystack: Benchmarking Multimodal Retrieval and Reasoning over 40K Images, Videos, and Documents

arXiv: https://arxiv.org/abs/2603.05697 · HTML: https://mlsanity.com/paper/2603.05697/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 53 of its 58 references that are in the map.
- First posted on arXiv 5 Mar 2026.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [Document understanding and multimodal RAG](https://mlsanity.com/mllm/document-understanding-and-multimodal-rag/)

## Builds on (most cited first, 30 of 53)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering](https://mlsanity.com/paper/1612.00837/) (2016)
- [MMBench: Is Your Multi-modal Model an All-around Player?](https://mlsanity.com/paper/2307.06281/) (2023)
- [InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models](https://mlsanity.com/paper/2504.10479/) (2025)
- [Towards VQA Models That Can Read](https://mlsanity.com/paper/1904.08920/) (2019)
- [Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Models](https://mlsanity.com/paper/1505.04870/) (2015)
- [DocVQA: A Dataset for VQA on Document Images](https://mlsanity.com/paper/2007.00398/) (2020)
- [SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features](https://mlsanity.com/paper/2502.14786/) (2025)
- [OK-VQA: A Visual Question Answering Benchmark Requiring External Knowledge](https://mlsanity.com/paper/1906.00067/) (2019)
- [MVBench: A Comprehensive Multi-modal Video Understanding Benchmark](https://mlsanity.com/paper/2311.17005/) (2023)
- [A-OKVQA: A Benchmark for Visual Question Answering using World Knowledge](https://mlsanity.com/paper/2206.01718/) (2022)
- [EgoSchema: A Diagnostic Benchmark for Very Long-form Video Language Understanding](https://mlsanity.com/paper/2308.09126/) (2023)
- [LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding](https://mlsanity.com/paper/2407.15754/) (2024)
- [MiniGPT-v2: large language model as a unified interface for vision-language multi-task learning](https://mlsanity.com/paper/2310.09478/) (2023)
- [VisRAG: Vision-based Retrieval-augmented Generation on Multi-modality Documents](https://mlsanity.com/paper/2410.10594/) (2024)
- [Can Pre-trained Vision and Language Models Answer Visual Information-Seeking Questions?](https://mlsanity.com/paper/2302.11713/) (2023)
- [MMLongBench-Doc: Benchmarking Long-context Document Understanding with Visualizations](https://mlsanity.com/paper/2407.01523/) (2024)
- [MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding](https://mlsanity.com/paper/2406.14515/) (2024)
- [MM-Embed: Universal Multimodal Retrieval with Multimodal LLMs](https://mlsanity.com/paper/2411.02571/) (2024)
- [E5-V: Universal Embeddings with Multimodal Large Language Models](https://mlsanity.com/paper/2407.12580/) (2024)
- [WebQA: Multihop and Multimodal QA](https://mlsanity.com/paper/2109.00590/) (2021)
- [OmniBench: Towards The Future of Universal Omni-Language Models](https://mlsanity.com/paper/2409.15272/) (2024)
- [MMSearch: Benchmarking the Potential of Large Models as Multi-modal Search Engines](https://mlsanity.com/paper/2409.12959/) (2024)
- [Ola: Pushing the Frontiers of Omni-Modal Language Model](https://mlsanity.com/paper/2502.04328/) (2025)
- [VideoVista: A Versatile Benchmark for Video Understanding and Reasoning](https://mlsanity.com/paper/2406.11303/) (2024)
- [MMIU: Multimodal Multi-image Understanding for Evaluating Large Vision-Language Models](https://mlsanity.com/paper/2408.02718/) (2024)
- [LongVT: Incentivizing "Thinking with Long Videos" via Native Tool Calling](https://mlsanity.com/paper/2511.20785/) (2025)

## Built on by (most cited first, 3 of 3)

- [SMMBench: A Benchmark for Source-Distributed Multimodal Agent Memory](https://mlsanity.com/paper/2605.15710/) (2026)
- [Personalizing Embodied Multimodal Large Language Model Agents over Long-term User Interactions](https://mlsanity.com/paper/2605.26256/) (2026)
- [OmniCapBench: A Deep-Structured Evaluation Framework for Fine-Grained Audio-Visual Captioning](https://mlsanity.com/paper/2610.12458/) (2026)

## Abstract (arXiv)

Multimodal large language models (MLLMs) achieve strong performance on benchmarks that evaluate text, image, or video understanding separately. However, these settings do not assess a critical real-world requirement, which involves retrieving relevant evidence from large, heterogeneous multimodal corpora prior to reasoning. Most existing benchmarks restrict retrieval to small, single-modality candidate sets, substantially simplifying the search space and overstating end-to-end reliability. To address this gap, we introduce MultiHaystack, the first benchmark designed to evaluate both retrieval and reasoning under large-scale, cross-modal conditions. MultiHaystack comprises over 46,000 multimodal retrieval candidates across documents, images, and videos, along with 747 open yet verifiable questions. Each question is grounded in a unique validated evidence item within the retrieval pool, requiring evidence localization across modalities and fine-grained reasoning. In our study, we find that models perform competitively when provided with the corresponding evidence, but their performance drops sharply when required to retrieve that evidence from the full corpus. Additionally, even the strongest retriever, E5-V, achieves only 40.8% Recall@1, while state-of-the-art MLLMs such as GPT-5 experience a significant drop in reasoning accuracy from 80.86% when provided with the corresponding evidence to 51.4% under top-5 retrieval. These results indicate that multimodal retrieval over heterogeneous pools remains a primary bottleneck for MLLMs, positioning MultiHaystack as a valuable testbed that highlights underlying limitations obscured by small-scale evaluations and promotes retrieval-centric advances in multimodal systems.