![[2602.16747] LiveClin: A Live Clinical Benchmark without Leakage](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.16747] LiveClin: A Live Clinical Benchmark without Leakage
Summary
LiveClin introduces a novel clinical benchmark for evaluating medical LLMs, addressing issues of data contamination and knowledge obsolescence with real-world case scenarios.
Why It Matters
As medical AI continues to evolve, ensuring the reliability of evaluations is crucial. LiveClin's approach offers a dynamic solution that reflects current clinical practices, potentially improving AI's real-world applicability in healthcare settings.
Key Takeaways
- LiveClin provides a continuously updated benchmark for medical LLMs.
- The benchmark is based on real-world clinical cases, enhancing relevance.
- Current models struggle with accuracy, highlighting the complexity of clinical scenarios.
Computer Science > Machine Learning arXiv:2602.16747 (cs) [Submitted on 18 Feb 2026] Title:LiveClin: A Live Clinical Benchmark without Leakage Authors:Xidong Wang, Shuqi Guo, Yue Shen, Junying Chen, Jian Wang, Jinjie Gu, Ping Zhang, Lei Liu, Benyou Wang View a PDF of the paper titled LiveClin: A Live Clinical Benchmark without Leakage, by Xidong Wang and 8 other authors View PDF HTML (experimental) Abstract:The reliability of medical LLM evaluation is critically undermined by data contamination and knowledge obsolescence, leading to inflated scores on static benchmarks. To address these challenges, we introduce LiveClin, a live benchmark designed for approximating real-world clinical practice. Built from contemporary, peer-reviewed case reports and updated biannually, LiveClin ensures clinical currency and resists data contamination. Using a verified AI-human workflow involving 239 physicians, we transform authentic patient cases into complex, multimodal evaluation scenarios that span the entire clinical pathway. The benchmark currently comprises 1,407 case reports and 6,605 questions. Our evaluation of 26 models on LiveClin reveals the profound difficulty of these real-world scenarios, with the top-performing model achieving a Case Accuracy of just 35.7%. In benchmarking against human experts, Chief Physicians achieved the highest accuracy, followed closely by Attending Physicians, with both surpassing most models. LiveClin thus provides a continuously evolving, clinicall...