OpenCompass 发布 v0.5.5：新增多领域评测基准并重构文档与 CI
原标题：0.5.5
AI 摘要
OpenCompass 发布 v0.5.5 版本，新增 CritPt、General365、AA-OmniScience、AA-LCR v1.1、HLE-Diamond、arena-hard-v2、ReasonZoo、MaritimeBench、WMT19 等覆盖物理、推理、长上下文、对齐、幻觉、生物医学、翻译和多模态评测的基准，并通过 VLMEvalKit 引入 ChartQAPro、MicroVQA 等数据集。该版本重构了中英文文档与 CI，移除 Python 3.8 支持并修复 Python 3.9 兼容性，同时修复了 OpenAI 系列模型的推理解析、数据集加载与下载等问题。
正文节选
# OpenCompass v0.5.5 Release Notes ## 🌟 Highlights ✨ **Integrated More Benchmarks:** Integrated new benchmarks spanning physics, reasoning, long-context, alignment, hallucination, maritime, biomedical, translation and multimodal evaluation, including CritPt, General365, AA-OmniScience v1.0 and v1.1, AA-LCR v1.1, HLE-Diamond, arena-hard-v2, ReasonZoo, LG-Bench, CHHallu-Src, GrandPhysics, PromptCBLUE Life Science, MaritimeBench, WMT19, LogiQA and ethics, plus ChartQAPro, MicroVQA, SFE and Sc