MatSciBench: Benchmarking the Reasoning Ability of Large Language Models in Materials Science
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911209159131136 |
|---|---|
| author | Zhang, Junkai Gan, Jingru Wang, Xiaoxuan Jia, Zian Gu, Changquan Chen, Jianpeng Zhu, Yanqiao Ma, Mingyu Derek Zhou, Dawei Li, Ling Wang, Wei |
| author_facet | Zhang, Junkai Gan, Jingru Wang, Xiaoxuan Jia, Zian Gu, Changquan Chen, Jianpeng Zhu, Yanqiao Ma, Mingyu Derek Zhou, Dawei Li, Ling Wang, Wei |
| contents | Large Language Models (LLMs) have demonstrated remarkable abilities in scientific reasoning, yet their reasoning capabilities in materials science remain underexplored. To fill this gap, we introduce MatSciBench, a comprehensive college-level benchmark comprising 1,340 problems that span the essential subdisciplines of materials science. MatSciBench features a structured and fine-grained taxonomy that categorizes materials science questions into 6 primary fields and 31 sub-fields, and includes a three-tier difficulty classification based on the reasoning length required to solve each question. MatSciBench provides detailed reference solutions enabling precise error analysis and incorporates multimodal reasoning through visual contexts in numerous questions. Evaluations of leading models reveal that even the highest-performing model, Gemini-2.5-Pro, achieves under 80% accuracy on college-level materials science questions, highlighting the complexity of MatSciBench. Our systematic analysis of different reasoning strategie--basic chain-of-thought, tool augmentation, and self-correction--demonstrates that no single method consistently excels across all scenarios. We further analyze performance by difficulty level, examine trade-offs between efficiency and accuracy, highlight the challenges inherent in multimodal reasoning tasks, analyze failure modes across LLMs and reasoning methods, and evaluate the influence of retrieval-augmented generation. MatSciBench thus establishes a comprehensive and solid benchmark for assessing and driving improvements in the scientific reasoning capabilities of LLMs within the materials science domain. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_12171 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MatSciBench: Benchmarking the Reasoning Ability of Large Language Models in Materials Science Zhang, Junkai Gan, Jingru Wang, Xiaoxuan Jia, Zian Gu, Changquan Chen, Jianpeng Zhu, Yanqiao Ma, Mingyu Derek Zhou, Dawei Li, Ling Wang, Wei Artificial Intelligence Large Language Models (LLMs) have demonstrated remarkable abilities in scientific reasoning, yet their reasoning capabilities in materials science remain underexplored. To fill this gap, we introduce MatSciBench, a comprehensive college-level benchmark comprising 1,340 problems that span the essential subdisciplines of materials science. MatSciBench features a structured and fine-grained taxonomy that categorizes materials science questions into 6 primary fields and 31 sub-fields, and includes a three-tier difficulty classification based on the reasoning length required to solve each question. MatSciBench provides detailed reference solutions enabling precise error analysis and incorporates multimodal reasoning through visual contexts in numerous questions. Evaluations of leading models reveal that even the highest-performing model, Gemini-2.5-Pro, achieves under 80% accuracy on college-level materials science questions, highlighting the complexity of MatSciBench. Our systematic analysis of different reasoning strategie--basic chain-of-thought, tool augmentation, and self-correction--demonstrates that no single method consistently excels across all scenarios. We further analyze performance by difficulty level, examine trade-offs between efficiency and accuracy, highlight the challenges inherent in multimodal reasoning tasks, analyze failure modes across LLMs and reasoning methods, and evaluate the influence of retrieval-augmented generation. MatSciBench thus establishes a comprehensive and solid benchmark for assessing and driving improvements in the scientific reasoning capabilities of LLMs within the materials science domain. |
| title | MatSciBench: Benchmarking the Reasoning Ability of Large Language Models in Materials Science |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2510.12171 |