Model Releases
Evaluation is All You Need: Strategic Overclaiming of LLM Reasoning Capabilities Through Evaluation Design
arXiv:2506.04734v3 Announce Type: replace Abstract: Reasoning models represented by the Deepseek-R1-Distill series have been widely adopted by the open-source community due to their strong performance
arXiv:2506.04734v3 Announce Type: replace Abstract: Reasoning models represented by the Deepseek-R1-Distill series have been widely adopted by the open-source community due to their strong performance in mathematics, science, programming, and other domains. However, our study reveals that their benchmark evaluation results are subject to significant fluctuations caused by various factors. Subtle differences in evaluation conditions can lead to substantial variations in results. Similar phenomena are observed in other open-source inference models fine-tuned based on the Deepseek-R1-Distill series, as well as in the QwQ-32B model, making their claimed performance improvements difficult to reproduce reliably. Therefore, we advocate for the establishment of a more rigorous paradigm for model performance evaluation and present our empirical assessments of the Deepseek-R1-Distill series models.
Related
- Do MLLMs Really Understand Space? A Mathematical Reasoning Evaluation
- MatSciBench: Benchmarking the Reasoning Ability of Large Language Models in Materials Science
- TRACE: Toulmin-based Reasoning Assessment through Constructive Elements for LLM CoT Evaluation
- Do VLMs Reason Like Engineers? A Benchmark and a Stage-wise Evaluation
Source: arXiv cs.AI | 2026-09-28