arXiv cs.AI· Jing Wang, Yitong Yang, Hao Dou, Yaomin Wu, Shuting He, Tingfei Shen, Jia Wu, Jiaqi Li, Yinglin Wang·· 12 小时前AI 评分31
TriGeoBench：包含图形与 TikZ 的多模态数学几何基准
TriGeoBench: A Multimodal Mathematical Geometry Benchmark with Figures and TikZ
AI 导读
TriGeoBench 是一个面向几何的多模态基准，包含 7,000 个问答对和 12,235 个带 TikZ 代码的几何图形，覆盖平面几何、立体几何、解析几何、函数与向量几何。每个实例由 {问答对、图像、TikZ} 三元组构成，问答对以 LaTeX 存储，图像为 RGB 图形，TikZ 保留几何拓扑与符号标注。对代表性开源与闭源 MLLM 的评测显示，模型在细粒度结构和图文对齐上仍存在持续挑战。
正文 · 原文
Abstract:Multimodal Large Language Models show strong vision-language capabilities, yet their geometry-centric reasoning remains insufficiently understood. Existing mathematical benchmarks typically rely on rasterized figures and cover limited geometric knowledge areas and response formats, making it difficult to assess whether models reason over geometric structures or exploit superficial visual patterns. We introduce TriGeoBench, a geometry-centric multimodal benchmark constructed from Chinese middle- and high-school mathematics examinations. TriGeoBench contains 7,000 question-answer pairs and 12,235 geometric figures with TikZ code, spanning plane geometry, solid geometry, analytic geometry, function, and vector geometry across diverse response formats. Each instance forms a {question-answer pair, image, TikZ} triplet: question-answer pairs are stored in LaTeX, images as RGB figures, and TikZ preserves geometric topology, symbolic annotations, and relational structure. We evaluate representative open- and closed-source MLLMs through image-to-TikZ and multimodal reasoning probes. Results reveal persistent fine-grained structure and image-text alignment challenges. Oracle experiments with human-curated reference TikZ suggest that accurate executable representations provide complementary geometric cues; they do not evaluate an end-to-end conversion pipeline. TriGeoBench offers a testbed for structure-aware mathematical reasoning.
| Comments: | 18 pages, 7 figures, 5 tables, including appendices. The first four authors contributed equally |
| Subjects: | Artificial Intelligence (cs.AI); Databases (cs.DB) |
| Cite as: | arXiv:2610.27626 [cs.AI] |
| (or arXiv:2610.27626v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.27626 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jing Wang [view email]
[v1]
Tue, 20 Oct 2026 04:52:30 UTC (4,918 KB)
来源：arXiv cs.AI · arxiv.org