Scaling Agentic Verifier for Competitive Coding
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912875406163968 |
|---|---|
| author | Ma, Zeyao Zhang, Jing Zhang, Xiaokang Yang, Jiaxi Zhang, Zongmeng Zhang, Jiajun Jing, Yuheng Zhang, Lei Zheng, Hao Zhao, Wenting Lin, Junyang Hui, Binyuan |
| author_facet | Ma, Zeyao Zhang, Jing Zhang, Xiaokang Yang, Jiaxi Zhang, Zongmeng Zhang, Jiajun Jing, Yuheng Zhang, Lei Zheng, Hao Zhao, Wenting Lin, Junyang Hui, Binyuan |
| contents | Large language models (LLMs) have demonstrated strong coding capabilities but still struggle to solve competitive programming problems correctly in a single attempt. Execution-based re-ranking offers a promising test-time scaling strategy, yet existing methods are constrained by either difficult test case generation or inefficient random input sampling. To address this limitation, we propose Agentic Verifier, an execution-based agent that actively reasons about program behaviors and searches for highly discriminative test inputs that expose behavioral discrepancies among candidate solutions. Through multi-turn interaction with code execution environments, the verifier iteratively refines the candidate input generator and produces targeted counterexamples rather than blindly sampling inputs. We train the verifier to acquire this discriminative input generation capability via a scalable pipeline combining large-scale data synthesis, rejection fine-tuning, and agentic reinforcement learning. Extensive experiments across five competitive programming benchmarks demonstrate consistent improvements over strong execution-based baselines, achieving up to +10-15% absolute gains in Best@K accuracy. Further analysis reveals clear test-time scaling behavior and highlights the verifier's broader potential beyond reranking. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_04254 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Scaling Agentic Verifier for Competitive Coding Ma, Zeyao Zhang, Jing Zhang, Xiaokang Yang, Jiaxi Zhang, Zongmeng Zhang, Jiajun Jing, Yuheng Zhang, Lei Zheng, Hao Zhao, Wenting Lin, Junyang Hui, Binyuan Computation and Language Large language models (LLMs) have demonstrated strong coding capabilities but still struggle to solve competitive programming problems correctly in a single attempt. Execution-based re-ranking offers a promising test-time scaling strategy, yet existing methods are constrained by either difficult test case generation or inefficient random input sampling. To address this limitation, we propose Agentic Verifier, an execution-based agent that actively reasons about program behaviors and searches for highly discriminative test inputs that expose behavioral discrepancies among candidate solutions. Through multi-turn interaction with code execution environments, the verifier iteratively refines the candidate input generator and produces targeted counterexamples rather than blindly sampling inputs. We train the verifier to acquire this discriminative input generation capability via a scalable pipeline combining large-scale data synthesis, rejection fine-tuning, and agentic reinforcement learning. Extensive experiments across five competitive programming benchmarks demonstrate consistent improvements over strong execution-based baselines, achieving up to +10-15% absolute gains in Best@K accuracy. Further analysis reveals clear test-time scaling behavior and highlights the verifier's broader potential beyond reranking. |
| title | Scaling Agentic Verifier for Competitive Coding |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2602.04254 |