BeyondBench: Contamination-Resistant Evaluation of Reasoning in Language Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908866334162944 |
|---|---|
| author | Srivastava, Gaurav Hussain, Aafiya Bi, Zhenyu Roy, Swastik Pitre, Priya Lu, Meng Ziyadi, Morteza Wang, Xuan |
| author_facet | Srivastava, Gaurav Hussain, Aafiya Bi, Zhenyu Roy, Swastik Pitre, Priya Lu, Meng Ziyadi, Morteza Wang, Xuan |
| contents | Evaluating language models fairly is increasingly difficult as static benchmarks risk contamination by training data, obscuring whether models truly reason or recall. We introduce BeyondBench, an evaluation framework using algorithmic problem generation to create mathematically grounded problems on the fly, ensuring each test remains uncontaminated. Our framework covers 44 algorithmic tasks with 117 variations across three difficulty levels: the Easy Suite (29 tasks) for arithmetic and statistics, the Medium Suite (5 tasks, 49 variations) for sequence patterns and reasoning, and the Hard Suite (10 tasks, 68 variations) for NP-complete and constraint satisfaction problems. Each task draws from a space exceeding 10^15 unique instances, with deterministically verified solutions. We evaluated 101 language models (85 open-source, 16 closed-source), spanning 0.5B to 141B parameters and multiple quantization schemes, using three-fold evaluation for robustness. Results reveal consistent reasoning deficiencies, with performance degrading sharply as complexity increases. In Hard Suite evaluations, Gemini-2.5-pro, Llama-3.3-70B, and Qwen2.5-72B achieved accuracies of 56.21%, 27.16%, and 33.37% respectively. Performance drops significantly without tool usage, with GPT-5, GPT-5-mini, and GPT-5-nano showing declines of 16.81%, 15.86%, and 43.95% in overall accuracy. Contamination resistance rests on three guarantees: (i) the problem space vastly exceeds any static dataset, (ii) every instance has a deterministically verifiable solution, and (iii) isomorphic transformations yield semantically equivalent but syntactically novel problems. BeyondBench redefines reasoning evaluation via genuine algorithmic problem-solving. Our leaderboard is at https://ctrl-gaurav.github.io/BeyondBench/, Python package at https://pypi.org/project/beyondbench/, and codebase at https://github.com/ctrl-gaurav/BeyondBench. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_24210 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | BeyondBench: Contamination-Resistant Evaluation of Reasoning in Language Models Srivastava, Gaurav Hussain, Aafiya Bi, Zhenyu Roy, Swastik Pitre, Priya Lu, Meng Ziyadi, Morteza Wang, Xuan Computation and Language Artificial Intelligence Machine Learning Evaluating language models fairly is increasingly difficult as static benchmarks risk contamination by training data, obscuring whether models truly reason or recall. We introduce BeyondBench, an evaluation framework using algorithmic problem generation to create mathematically grounded problems on the fly, ensuring each test remains uncontaminated. Our framework covers 44 algorithmic tasks with 117 variations across three difficulty levels: the Easy Suite (29 tasks) for arithmetic and statistics, the Medium Suite (5 tasks, 49 variations) for sequence patterns and reasoning, and the Hard Suite (10 tasks, 68 variations) for NP-complete and constraint satisfaction problems. Each task draws from a space exceeding 10^15 unique instances, with deterministically verified solutions. We evaluated 101 language models (85 open-source, 16 closed-source), spanning 0.5B to 141B parameters and multiple quantization schemes, using three-fold evaluation for robustness. Results reveal consistent reasoning deficiencies, with performance degrading sharply as complexity increases. In Hard Suite evaluations, Gemini-2.5-pro, Llama-3.3-70B, and Qwen2.5-72B achieved accuracies of 56.21%, 27.16%, and 33.37% respectively. Performance drops significantly without tool usage, with GPT-5, GPT-5-mini, and GPT-5-nano showing declines of 16.81%, 15.86%, and 43.95% in overall accuracy. Contamination resistance rests on three guarantees: (i) the problem space vastly exceeds any static dataset, (ii) every instance has a deterministically verifiable solution, and (iii) isomorphic transformations yield semantically equivalent but syntactically novel problems. BeyondBench redefines reasoning evaluation via genuine algorithmic problem-solving. Our leaderboard is at https://ctrl-gaurav.github.io/BeyondBench/, Python package at https://pypi.org/project/beyondbench/, and codebase at https://github.com/ctrl-gaurav/BeyondBench. |
| title | BeyondBench: Contamination-Resistant Evaluation of Reasoning in Language Models |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2509.24210 |