# Migician: Revealing the Magic of Free-Form Multi-Image Grounding in Multimodal Large Language Models

arXiv: https://arxiv.org/abs/2501.05767 · HTML: https://mlsanity.com/paper/2501.05767/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 45 of its 53 references that are in the map.
- First posted on arXiv 10 Jan 2025.

## Placed in

- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)

## Builds on (most cited first, 30 of 45)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations](https://mlsanity.com/paper/1602.07332/) (2016)
- [Qwen2 Technical Report](https://mlsanity.com/paper/2407.10671/) (2024)
- [MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models](https://mlsanity.com/paper/2306.13394/) (2023)
- [Ego4D: Around the World in 3,000 Hours of Egocentric Video](https://mlsanity.com/paper/2110.07058/) (2021)
- [Editing Models with Task Arithmetic](https://mlsanity.com/paper/2212.04089/) (2022)
- [Kosmos-2: Grounding Multimodal Large Language Models to the World](https://mlsanity.com/paper/2306.14824/) (2023)
- [MiniCPM-V: A GPT-4V Level MLLM on Your Phone](https://mlsanity.com/paper/2408.01800/) (2024)
- [Shikra: Unleashing Multimodal LLM's Referential Dialogue Magic](https://mlsanity.com/paper/2306.15195/) (2023)
- [MathQA: Towards Interpretable Math Word Problem Solving with Operation-Based Formalisms](https://mlsanity.com/paper/1905.13319/) (2019)
- [MagicBrush: A Manually Annotated Dataset for Instruction-Guided Image Editing](https://mlsanity.com/paper/2306.10012/) (2023)
- [LLaVA-NeXT-Interleave: Tackling Multi-image, Video, and 3D in Large Multimodal Models](https://mlsanity.com/paper/2407.07895/) (2024)
- [BLINK: Multimodal Large Language Models Can See but Not Perceive](https://mlsanity.com/paper/2404.12390/) (2024)
- [MOT16: A Benchmark for Multi-Object Tracking](https://mlsanity.com/paper/1603.00831/) (2016)
- [Ferret: Refer and Ground Anything Anywhere at Any Granularity](https://mlsanity.com/paper/2310.07704/) (2023)
- [V*: Guided Visual Search as a Core Mechanism in Multimodal LLMs](https://mlsanity.com/paper/2312.14135/) (2023)
- [GOT-10k: A Large High-Diversity Benchmark for Generic Object Tracking in the Wild](https://mlsanity.com/paper/1810.11981/) (2018)
- [GLaMM: Pixel Grounding Large Multimodal Model](https://mlsanity.com/paper/2311.03356/) (2023)
- [LaSOT: A High-quality Benchmark for Large-scale Single Object Tracking](https://mlsanity.com/paper/1809.07845/) (2018)
- [InternLM2 Technical Report](https://mlsanity.com/paper/2403.17297/) (2024)
- [TrackingNet: A Large-Scale Dataset and Benchmark for Object Tracking in the Wild](https://mlsanity.com/paper/1803.10794/) (2018)
- [Driving into the Future: Multiview Visual Forecasting and Planning with World Model for Autonomous Driving](https://mlsanity.com/paper/2311.17918/) (2023)
- [GPT4RoI: Instruction Tuning Large Language Model on Region-of-Interest](https://mlsanity.com/paper/2307.03601/) (2023)
- [mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Models](https://mlsanity.com/paper/2408.04840/) (2024)
- [IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning](https://mlsanity.com/paper/2110.13214/) (2021)
- [mPLUG-DocOwl 1.5: Unified Structure Learning for OCR-free Document Understanding](https://mlsanity.com/paper/2403.12895/) (2024)

## Built on by (most cited first, 22 of 22)

- [Multimodal Chain-of-Thought Reasoning: A Comprehensive Survey](https://mlsanity.com/paper/2503.12605/) (2025)
- [UniVG-R1: Reasoning Guided Universal Visual Grounding with Reinforcement Learning](https://mlsanity.com/paper/2505.14231/) (2025)
- [Thinking with Visual Abstract: Enhancing Multimodal Reasoning via Visual Abstraction](https://mlsanity.com/paper/2505.20164/) (2025)
- [Unveiling the Potential of Vision-Language-Action Models with Open-Ended Multimodal Instructions](https://mlsanity.com/paper/2505.11214/) (2025)
- [Imagination Helps Visual Reasoning, But Not Yet in Latent Space](https://mlsanity.com/paper/2602.22766/) (2026)
- [PeRL: Permutation-Enhanced Reinforcement Learning for Interleaved Vision-Language Reasoning](https://mlsanity.com/paper/2506.14907/) (2025)
- [Improving the Reasoning of Multi-Image Grounding in MLLMs via Reinforcement Learning](https://mlsanity.com/paper/2507.00748/) (2025)
- [MedSG-Bench: A Benchmark for Medical Image Sequences Grounding](https://mlsanity.com/paper/2505.11852/) (2025)
- [VP-Bench: A Comprehensive Benchmark for Visual Prompting in Multimodal Large Language Models](https://mlsanity.com/paper/2511.11438/) (2025)
- [Training Multi-Image Vision Agents via End2End Reinforcement Learning](https://mlsanity.com/paper/2512.08980/) (2025)
- [MIRG-RL: Multi-Image Reasoning and Grounding with Reinforcement Learning](https://mlsanity.com/paper/2509.21788/) (2025)
- [Mixture-of-Thought-Tokens: Unifying Perception and Reasoning for Free-form Multimodal Grounding](https://mlsanity.com/paper/2607.24407/) (2026)
- [RefBench-PRO: Perceptual and Reasoning Oriented Benchmark for Referring Expression Comprehension](https://mlsanity.com/paper/2512.06276/) (2025)
- [RGBX-R1: Visual Modality Chain-of-Thought Guided Reinforcement Learning for Multimodal Grounding](https://mlsanity.com/paper/2602.00504/) (2026)
- [GroundedSurg: A Multi-Procedure Benchmark for Language-Conditioned Surgical Tool Segmentation](https://mlsanity.com/paper/2603.01108/) (2026)
- [CGC: Compositional Grounded Contrast for Fine-Grained Multi-Image Understanding](https://mlsanity.com/paper/2604.22498/) (2026)
- [Does Seeing More Mean Knowing More? Mono-Anchored Advantage Normalization for Multi-Source Visual Reasoning](https://mlsanity.com/paper/2605.25437/) (2026)
- [DifferAD-R1: A Difference-Guided IndustrialAnomaly Localization with Multimodal LargeLanguage Models](https://mlsanity.com/paper/2606.16601/) (2026)
- [From Structure to Synergy: A Survey of Vision-Language Perception Paradigm Evolution in Multimodal Large Language Models](https://mlsanity.com/paper/2606.26196/) (2026)
- [DiCoBench: Benchmarking Multi-Image Fine-Grained Perception via Differential and Commonality Visual Cues](https://mlsanity.com/paper/2606.26602/) (2026)
- [MinCU: A Fine-Grained Benchmark for Grounded Minimal-Change Understanding in Image Pairs](https://mlsanity.com/paper/2609.23336/) (2026)
- [Rho: A Foundation for Efficiently Adaptable VLA Models](https://mlsanity.com/paper/2609.38164/) (2026)

## Abstract (arXiv)

The recent advancement of Multimodal Large Language Models (MLLMs) has significantly improved their fine-grained perception of single images and general comprehension across multiple images. However, existing MLLMs still face challenges in achieving precise grounding in complex multi-image scenarios. To address this, we first explore a Chain-of-Thought (CoT) framework that integrates single-image grounding with multi-image comprehension. While partially effective, it remains unstable and struggles to capture abstract visual information due to its non-end-to-end nature. Therefore, we introduce Migician, the first multi-image grounding model capable of performing free-form and accurate grounding across multiple images. To support this, we present the MGrounding-630k dataset, which comprises data for several multi-image grounding tasks derived from existing datasets, along with newly generated free-form grounding instruction-following data. Furthermore, we propose MIG-Bench, a comprehensive benchmark specifically designed for evaluating multi-image grounding capabilities. Experimental results demonstrate that our model achieves significantly superior multi-image grounding capabilities, outperforming the best existing MLLMs by 24.94% and even surpassing much larger 70B models. Our code, model, dataset, and benchmark are fully open-sourced at https://migician-vg.github.io/.