Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.27959 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911557548507136 |
|---|---|
| author | Liu, Ruiyao Shen, Hui Zhang, Ping Hsieh, Yunta Zhang, Yifan Xu, Jing Chen, Sicheng Li, Junchen Lu, Jiawei Ma, Jianing Mo, Jiaqi Han, Qi Zhang, Zhen Wan, Zhongwei Xiong, Jing Wang, Xin Liu, Ziyuan Cao, Hangrui Wong, Ngai |
| author_facet | Liu, Ruiyao Shen, Hui Zhang, Ping Hsieh, Yunta Zhang, Yifan Xu, Jing Chen, Sicheng Li, Junchen Lu, Jiawei Ma, Jianing Mo, Jiaqi Han, Qi Zhang, Zhen Wan, Zhongwei Xiong, Jing Wang, Xin Liu, Ziyuan Cao, Hangrui Wong, Ngai |
| contents | Modern generative models have demonstrated the ability to solve challenging mathematical problems. In many real-world settings, however, mathematical solutions must be expressed visually through diagrams, plots, geometric constructions, and structured symbolic layouts, where correctness depends on precise visual composition. This naturally raises the question of whether generative models can still do so when the answer must be rendered visually rather than written in text? To study this problem, we introduce MathGen, a rigorous benchmark of 900 problems spanning seven core domains, each paired with an executable verifier under a Script-as-a-Judge protocol for deterministic and objective evaluation. Experiments on representative open-source and proprietary text-to-image models show that mathematical fidelity remains a major bottleneck: even the best closed-source model reaches only 42.0% overall accuracy, while open-source models achieve just ~ 1-11%, often near 0% on structured tasks. Overall, current T2I models remain far from competent at even elementary mathematical visual generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_27959 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MathGen: Revealing the Illusion of Mathematical Competence through Text-to-Image Generation Liu, Ruiyao Shen, Hui Zhang, Ping Hsieh, Yunta Zhang, Yifan Xu, Jing Chen, Sicheng Li, Junchen Lu, Jiawei Ma, Jianing Mo, Jiaqi Han, Qi Zhang, Zhen Wan, Zhongwei Xiong, Jing Wang, Xin Liu, Ziyuan Cao, Hangrui Wong, Ngai Computer Vision and Pattern Recognition Modern generative models have demonstrated the ability to solve challenging mathematical problems. In many real-world settings, however, mathematical solutions must be expressed visually through diagrams, plots, geometric constructions, and structured symbolic layouts, where correctness depends on precise visual composition. This naturally raises the question of whether generative models can still do so when the answer must be rendered visually rather than written in text? To study this problem, we introduce MathGen, a rigorous benchmark of 900 problems spanning seven core domains, each paired with an executable verifier under a Script-as-a-Judge protocol for deterministic and objective evaluation. Experiments on representative open-source and proprietary text-to-image models show that mathematical fidelity remains a major bottleneck: even the best closed-source model reaches only 42.0% overall accuracy, while open-source models achieve just ~ 1-11%, often near 0% on structured tasks. Overall, current T2I models remain far from competent at even elementary mathematical visual generation. |
| title | MathGen: Revealing the Illusion of Mathematical Competence through Text-to-Image Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.27959 |