Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.01241 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911188528398336 |
|---|---|
| author | Wei, Hu Xu, Ze Yang, Boyu Miao, Linlin Zhai, Weiqi Li, Yihan Li, Zixuan Wang, Zhijun Wang, Boya Yu, Jianwei Yuan, Jialing Zhang, Xiaoyue He, Cheng Chen, Minglei Zhang, Zifan Li, Qianhui Wang, Wei Xu, Xiang |
| author_facet | Wei, Hu Xu, Ze Yang, Boyu Miao, Linlin Zhai, Weiqi Li, Yihan Li, Zixuan Wang, Zhijun Wang, Boya Yu, Jianwei Yuan, Jialing Zhang, Xiaoyue He, Cheng Chen, Minglei Zhang, Zifan Li, Qianhui Wang, Wei Xu, Xiang |
| contents | Large language models (LLMs) now perform strongly on many public math suites, yet frontier separation within mathematics increasingly suffers from ceiling effects. We present two complementary benchmarks: SKYLENAGE-ReasoningMATH, a 100-item, structure-aware diagnostic set with per-item metadata on length, numeric density, and symbolic complexity; and SKYLENAGE-MATH, a 150-item contest-style suite spanning four stages from high school to doctoral under a seven-subject taxonomy. We evaluate fifteen contemporary LLM variants under a single setup and analyze subject x model and grade x model performance. On the contest suite, the strongest model reaches 44% while the runner-up reaches 37%; accuracy declines from high school to doctoral, and top systems exhibit a doctoral-to-high-school retention near 79%. On the reasoning set, the best model attains 81% overall, and hardest-slice results reveal clear robustness gaps between leaders and the mid-tier. In summary, we release SKYLENAGE-ReasoningMATH and report aggregate results for SKYLENAGE-MATH; together, SKYLENAGE provides a hard, reasoning-centered and broadly covering math benchmark with calibrated difficulty and rich metadata, serving as a reference benchmark for future evaluations of mathematical reasoning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_01241 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SKYLENAGE Technical Report: Mathematical Reasoning and Contest-Innovation Benchmarks for Multi-Level Math Evaluation Wei, Hu Xu, Ze Yang, Boyu Miao, Linlin Zhai, Weiqi Li, Yihan Li, Zixuan Wang, Zhijun Wang, Boya Yu, Jianwei Yuan, Jialing Zhang, Xiaoyue He, Cheng Chen, Minglei Zhang, Zifan Li, Qianhui Wang, Wei Xu, Xiang Computation and Language Large language models (LLMs) now perform strongly on many public math suites, yet frontier separation within mathematics increasingly suffers from ceiling effects. We present two complementary benchmarks: SKYLENAGE-ReasoningMATH, a 100-item, structure-aware diagnostic set with per-item metadata on length, numeric density, and symbolic complexity; and SKYLENAGE-MATH, a 150-item contest-style suite spanning four stages from high school to doctoral under a seven-subject taxonomy. We evaluate fifteen contemporary LLM variants under a single setup and analyze subject x model and grade x model performance. On the contest suite, the strongest model reaches 44% while the runner-up reaches 37%; accuracy declines from high school to doctoral, and top systems exhibit a doctoral-to-high-school retention near 79%. On the reasoning set, the best model attains 81% overall, and hardest-slice results reveal clear robustness gaps between leaders and the mid-tier. In summary, we release SKYLENAGE-ReasoningMATH and report aggregate results for SKYLENAGE-MATH; together, SKYLENAGE provides a hard, reasoning-centered and broadly covering math benchmark with calibrated difficulty and rich metadata, serving as a reference benchmark for future evaluations of mathematical reasoning. |
| title | SKYLENAGE Technical Report: Mathematical Reasoning and Contest-Innovation Benchmarks for Multi-Level Math Evaluation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2510.01241 |