# TangramPuzzle: Evaluating Multimodal Large Language Models with Compositional Spatial Reasoning

arXiv: https://arxiv.org/abs/2601.16520 · HTML: https://mlsanity.com/paper/2601.16520/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 43 of its 103 references that are in the map.
- First posted on arXiv 23 Jan 2026.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)
- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)

## Builds on (most cited first, 30 of 43)

- [DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding](https://mlsanity.com/paper/2412.10302/) (2024)
- [MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark](https://mlsanity.com/paper/2409.02813/) (2024)
- [AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents](https://mlsanity.com/paper/2405.14573/) (2024)
- [ScreenSpot-Pro: GUI Grounding for Professional High-Resolution Computer Use](https://mlsanity.com/paper/2504.07981/) (2025)
- [LogicVista: Multimodal LLM Logical Reasoning Benchmark in Visual Contexts](https://mlsanity.com/paper/2407.04973/) (2024)
- [MANTIS: Interleaved Multi-Image Instruction Tuning](https://mlsanity.com/paper/2405.01483/) (2024)
- [Teaching CLIP to Count to Ten](https://mlsanity.com/paper/2302.12066/) (2023)
- [OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning](https://mlsanity.com/paper/2501.00321/) (2024)
- [Omni3D: A Large Benchmark and Model for 3D Object Detection in the Wild](https://mlsanity.com/paper/2207.10660/) (2022)
- [3DSRBench: A Comprehensive 3D Spatial Reasoning Benchmark](https://mlsanity.com/paper/2412.07825/) (2024)
- [Scaling Computer-Use Grounding via User Interface Decomposition and Synthesis](https://mlsanity.com/paper/2505.13227/) (2025)
- [ChartQAPro: A More Diverse and Challenging Benchmark for Chart Question Answering](https://mlsanity.com/paper/2504.05506/) (2025)
- [Benchmarking and Improving Detail Image Caption](https://mlsanity.com/paper/2405.19092/) (2024)
- [Natural Language Understanding and Inference with MLLM in Visual Question Answering: A Survey](https://mlsanity.com/paper/2411.17558/) (2024)
- [Are we ready for a new paradigm shift? A Survey on Visual Deep MLP](https://mlsanity.com/paper/2111.04060/) (2021)
- [The Past Mistake is the Future Wisdom: Error-driven Contrastive Probability Optimization for Chinese Spell Checking](https://mlsanity.com/paper/2203.00991/) (2022)
- [On the (In)Effectiveness of Large Language Models for Chinese Text Correction](https://mlsanity.com/paper/2307.09007/) (2023)
- [The Curse of Multi-Modalities: Evaluating Hallucinations of Large Multimodal Models across Language, Visual, and Audio](https://mlsanity.com/paper/2410.12787/) (2024)
- [LEGO-Puzzles: How Good Are MLLMs at Multi-Step Spatial Reasoning?](https://mlsanity.com/paper/2503.19990/) (2025)
- [Abstract Visual Reasoning with Tangram Shapes](https://mlsanity.com/paper/2211.16492/) (2022)
- [LatEval: An Interactive LLMs Evaluation Benchmark with Incomplete Information from Lateral Thinking Puzzles](https://mlsanity.com/paper/2308.10855/) (2023)
- [MC-Bench: A Benchmark for Multi-Context Visual Grounding in the Era of MLLMs](https://mlsanity.com/paper/2410.12332/) (2024)
- [SpatialBench: Benchmarking Multimodal Large Language Models for Spatial Cognition](https://mlsanity.com/paper/2511.21471/) (2025)
- [R-Bench: Graduate-level Multi-disciplinary Benchmarks for LLM & MLLM Complex Reasoning Evaluation](https://mlsanity.com/paper/2505.02018/) (2025)
- [Multimodal Spatial Reasoning in the Large Model Era: A Survey and Benchmarks](https://mlsanity.com/paper/2510.25760/) (2025)
- [ChartEdit: How Far Are MLLMs From Automating Chart Analysis? Evaluating MLLMs' Capability via Chart Editing](https://mlsanity.com/paper/2505.11935/) (2025)
- [Learning from the Dictionary: Heterogeneous Knowledge Guided Fine-tuning for Chinese Spell Checking](https://mlsanity.com/paper/2210.10320/) (2022)
- [Benchmarking Large Vision-Language Models via Directed Scene Graph for Comprehensive Image Captioning](https://mlsanity.com/paper/2412.08614/) (2024)
- [OS-Kairos: Adaptive Interaction for MLLM-Powered GUI Agents](https://mlsanity.com/paper/2503.16465/) (2025)
- [OCR-Reasoning Benchmark: Unveiling the True Capabilities of MLLMs in Complex Text-Rich Image Reasoning](https://mlsanity.com/paper/2505.17163/) (2025)

## Built on by (most cited first, 4 of 4)

- [Cognitive Mismatch in Multimodal Large Language Models for Discrete Symbol Understanding](https://mlsanity.com/paper/2603.18472/) (2026)
- [From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI](https://mlsanity.com/paper/2606.14502/) (2026)
- [TopoAgent: A Self-Evolving Topological Agent for Multimodal Scientific Reasoning](https://mlsanity.com/paper/2607.14658/) (2026)
- [INSPIRE: An Internalize-Then-Improve Approach for Example-Driven Mathematical Reasoning](https://mlsanity.com/paper/2608.27501/) (2026)

## Abstract (arXiv)

Multimodal Large Language Models (MLLMs) have achieved remarkable progress in visual recognition and semantic understanding, yet precise compositional spatial reasoning under geometric constraints remains underexplored. Existing benchmarks mainly assess coarse spatial relations and rarely support rigorous geometric verification or multiple valid solutions in constructive tasks. To address these limitations, we introduce TangramPuzzle, a benchmark comprising 668 validated configurations and 1,336 instances for evaluating compositional spatial reasoning under geometric constraints. We formulate the Tangram Construction Expression (TCE) to encode tangram configurations with machine-verifiable geometric specifications, together with a multi-solution-aware verifier that separately evaluates geometric validity and silhouette fidelity. TangramPuzzle covers two complementary evaluation tasks: part-to-whole Outline Prediction and whole-to-part End-to-End Assembly Generation. We conduct extensive evaluation experiments on advanced open-source and proprietary models, revealing an interesting insight: MLLMs tend to prioritize matching the target silhouette while neglecting geometric constraints, leading to distortions or deformations of the pieces.