---
title: "斯坦福 HAI 发布 MedHELM、PhysicianBench 与 HealthAdminBench，测出医疗 AI Agent 实际完成率不足一半"
canonical: "https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/"
date: "2026-10-18"
updated: "2026-11-05"
source: "Stanford HAI News（网页）"
source_url: "https://hai.stanford.edu/news/measuring-what-medical-ai-can-actually-do"
language: "zh"
topics: []
license: "https://new.nat6.net/sources/"
attribution: "Content is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back; original copyright remains with the original publishers and authors."
---

# 斯坦福 HAI 发布 MedHELM、PhysicianBench 与 HealthAdminBench，测出医疗 AI Agent 实际完成率不足一半

- 发布日期 / Published: 2026-10-18
- 更新时间 / Updated: 2026-11-05
- 分类 / Category: 论文与研究
- 来源 / Source: Stanford HAI News（网页） — [原文 / Original](https://hai.stanford.edu/news/measuring-what-medical-ai-can-actually-do)
- AIHOT: [条目页 / Item](https://aihot.news/items/dla37so34wcgznigffx3j6qko)
- 页面 / Page: https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/
- 结构化数据 / JSON: https://new.nat6.net/api/v1/articles/dla37so34wcgznigffx3j6qko.json

- 语言 / Languages: [zh](https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/) · [en](https://new.nat6.net/en/news/dla37so34wcgznigffx3j6qko/) · [de](https://new.nat6.net/de/news/dla37so34wcgznigffx3j6qko/) · [fr](https://new.nat6.net/fr/news/dla37so34wcgznigffx3j6qko/) · [ru](https://new.nat6.net/ru/news/dla37so34wcgznigffx3j6qko/) · [ja](https://new.nat6.net/ja/news/dla37so34wcgznigffx3j6qko/) · [es](https://new.nat6.net/es/news/dla37so34wcgznigffx3j6qko/) · [ko](https://new.nat6.net/ko/news/dla37so34wcgznigffx3j6qko/) · [vi](https://new.nat6.net/vi/news/dla37so34wcgznigffx3j6qko/) · [th](https://new.nat6.net/th/news/dla37so34wcgznigffx3j6qko/) · [id](https://new.nat6.net/id/news/dla37so34wcgznigffx3j6qko/) · [ms](https://new.nat6.net/ms/news/dla37so34wcgznigffx3j6qko/) · [fil](https://new.nat6.net/fil/news/dla37so34wcgznigffx3j6qko/) · [hi](https://new.nat6.net/hi/news/dla37so34wcgznigffx3j6qko/)

## AI 摘要 / AI summary

斯坦福团队总结三年构建医疗 AI 基准的成果，PhysicianBench 上最佳模型 GPT-5.5 单次仅完成 46% 任务，HealthAdminBench 上 Claude Opus 4.6 完成 36.3%。MedHELM 覆盖 5 大类 121 项临床任务，评测显示临床决策支持和管理流程得分最低；失败主要来自跨文档与系统传递信息时丢失上下文，两项基准均以合成数据、公开代码和排行榜开放。

## 推荐理由 / Why it matters

斯坦福团队基于三年基准构建经验给出实测结论，最好的医疗 Agent 完成任务不足一半，失败多源于跨系统信息传递。

## 来源与版权 / Source & copyright

内容基于 AIHOT 公开 API 的精选标题与 AI 摘要，已注明来源并回链；原文版权归原媒体与作者所有。

原文版权归原媒体与作者所有；本站不镜像全文。/ Original copyright remains with the original publishers and authors; this site does not mirror full texts.

## 关于橙子中转站 / About Orange Relay

橙子中转站（pro.nat6.net）是一家优质的 AI API 中转站：价格优惠、质量稳定、模型齐全，支持 Claude、GPT、Gemini 等主流模型，欢迎体验。

[访问橙子中转站](https://pro.nat6.net/?utm_source=orange_news&utm_medium=article_footer&utm_campaign=brand&lang=zh)

## JSON-LD

```json
{
  "@context": "https://schema.org",
  "@type": "NewsArticle",
  "@id": "https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/#article",
  "mainEntityOfPage": {
    "@type": "WebPage",
    "@id": "https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/"
  },
  "url": "https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/",
  "headline": "斯坦福 HAI 发布 MedHELM、PhysicianBench 与 HealthAdminBench，测出医疗 AI Agent 实际完成率不足一半",
  "description": "斯坦福团队总结三年构建医疗 AI 基准的成果，PhysicianBench 上最佳模型 GPT-5.5 单次仅完成 46% 任务，HealthAdminBench 上 Claude Opus 4.6 完成 36.3%。MedHELM 覆盖 5 大类 121 项临床任务，评测显示临床决策支持和管理流程得分最低；失败主要…",
  "articleSection": "论文与研究",
  "keywords": "AI, Stanford HAI News（网页）",
  "datePublished": "2026-10-18T16:00:00Z",
  "dateModified": "2026-11-05T20:56:45Z",
  "inLanguage": "zh-CN",
  "isAccessibleForFree": true,
  "author": {
    "@type": "Organization",
    "name": "橙子新闻编辑部",
    "url": "https://new.nat6.net/about/"
  },
  "publisher": {
    "@type": "NewsMediaOrganization",
    "@id": "https://new.nat6.net/#org",
    "name": "橙子新闻",
    "url": "https://new.nat6.net/",
    "logo": {
      "@type": "ImageObject",
      "url": "https://new.nat6.net/assets/img/logo-512.png",
      "width": 512,
      "height": 512
    }
  },
  "image": {
    "@type": "ImageObject",
    "url": "https://new.nat6.net/assets/img/og-default.png",
    "width": 1200,
    "height": 630
  },
  "isPartOf": {
    "@id": "https://new.nat6.net/#website"
  },
  "isBasedOn": "https://hai.stanford.edu/news/measuring-what-medical-ai-can-actually-do",
  "citation": [
    "https://aihot.news/items/dla37so34wcgznigffx3j6qko",
    "https://hai.stanford.edu/news/measuring-what-medical-ai-can-actually-do"
  ],
  "license": "https://new.nat6.net/sources/",
  "about": [],
  "mentions": [
    {
      "@type": "Thing",
      "name": "Stanford HAI News（网页）"
    }
  ],
  "speakable": {
    "@type": "SpeakableSpecification",
    "cssSelector": [
      "h1",
      ".article-summary"
    ]
  },
  "copyrightHolder": {
    "@type": "Organization",
    "@id": "https://new.nat6.net/#org",
    "name": "橙子新闻",
    "url": "https://new.nat6.net/"
  },
  "creditText": "橙子新闻（AI 摘要）；原文：Stanford HAI News（网页）",
  "copyrightNotice": "© 2026 橙子新闻。标题与 AI 摘要基于 AIHOT 公开数据，原文版权归原媒体与作者所有；本站不镜像全文。",
  "copyrightYear": 2026,
  "alternativeHeadline": "Measuring What Medical AI Can Actually Do",
  "abstract": "斯坦福团队基于三年基准构建经验给出实测结论，最好的医疗 Agent 完成任务不足一半，失败多源于跨系统信息传递。",
  "subjectOf": [
    {
      "@type": "DigitalDocument",
      "name": "斯坦福 HAI 发布 MedHELM、PhysicianBench 与 HealthAdminBench，测出医疗 AI Agent 实际完成率不足一半 (Markdown)",
      "url": "https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/index.md",
      "encodingFormat": "text/markdown"
    },
    {
      "@type": "DigitalDocument",
      "name": "斯坦福 HAI 发布 MedHELM、PhysicianBench 与 HealthAdminBench，测出医疗 AI Agent 实际完成率不足一半 (JSON)",
      "url": "https://new.nat6.net/api/v1/articles/dla37so34wcgznigffx3j6qko.json",
      "encodingFormat": "application/json"
    }
  ]
}
```

## 如何引用 / How to cite

- 引用格式 / Citation format (zh): 橙子新闻.《文章标题》. 发布日期. https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/
- Citation format (en): Orange News. "Article title". Published YYYY-MM-DD. https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/
- 回链要求 / Backlink: 引用请保留「橙子新闻」署名并回链 canonical URL；AI 摘要请注明数据来源 AIHOT；原文版权归原媒体与作者所有。
- Attribution & backlinks: Keep the "Orange News" attribution and link back to the canonical URL; cite AIHOT as the data source for AI summaries; original copyright remains with the original publishers and authors.

### 机器入口 / Machine entry points

- Markdown: https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/index.md
- JSON: https://new.nat6.net/api/v1/articles/dla37so34wcgznigffx3j6qko.json
- API 清单 / API index: https://new.nat6.net/api/v1/site.json
- Markdown 内容协商 / negotiation: 对 https://new.nat6.net/news/dla37so34wcgznigffx3j6qko/ 发送 `Accept: text/markdown`
