Ploba logo

Find the right AI tool for the job: apps, agents, APIs, MCP servers and datasets, explained plainly.

Get the AI Digest

Discover

  • Explore
  • Top Charts
  • Categories
  • Agents
  • MCP servers
  • Developer tools

Learn & community

  • Learn AI
  • AI Glossary
  • Blog
  • Hackathons
  • People
  • AI Digest

For makers

  • List your tool
  • Docs
  • Claim a listing
  • Advertise
  • Write for us

Company

  • About
  • Careers
  • Changelog
  • Trust
  • Security
  • Contact

© 2026 Ploba. All rights reserved.

  • Privacy
  • Terms
  • Cookies
  • Advertising
  • All policies
llms.txt
Ploba logo
DiscoverHackathonsPeopleLearnAI Digest
Sign inGet started
HomeTop ChartsExploreCategories
TypesWeb apps1,894Agents585Plugins416APIs376Mobile apps152Datasets85
AdvertisingYour product hereReach people browsing AI tools on Ploba.Advertise hereAdvertisingYour product hereReach people browsing AI tools on Ploba.Advertise here
  1. Explore
  2. /
  3. Health
  4. /
  5. HealthBench
Health
HealthBench icon

HealthBench

by OpenAI

HealthBench is an OpenAI benchmark of 5,000 realistic health conversations, each graded with a physician-written rubric. It measures how well AI models respond in health settings.

Dataset·Free

Visit openai.com(opens in a new tab)

About HealthBench

Key features of HealthBench

  • 5,000 realistic health conversations
  • Multi-turn and multilingual cases
  • Physician-written rubric per conversation
  • 48,562 unique rubric criteria
  • Seven themes such as emergency situations and global health
  • Axes such as accuracy and communication quality
  • Built with 262 physicians in 60 countries
  • Paper and code published

Who HealthBench is for

  • AI model developers
  • Health AI researchers
  • Safety and evaluation teams

How to use HealthBench

  1. 1

    Step 1: Read the paper and the code linked on the announcement page.

  2. 2

    Step 2: Run your model on the conversations and collect its last-turn responses.

  3. 3

    Step 3: Grade each response against the conversation's rubric criteria with a model-based grader.

  4. 4

    Step 4: Sum the points of the criteria met and divide by the maximum possible.

  5. 5

    Step 5: Break results down by theme and axis.

HealthBench FAQ

How are responses graded?

A model-based grader checks each physician-written rubric criterion, and the overall score is the points earned over the maximum possible.

What do the conversations look like?

They are multi-turn and multilingual, with layperson and clinician personas across many specialties, and were selected for difficulty.

Who wrote the rubrics?

A group of 262 physicians who have practiced in 60 countries, proficient in 49 languages and trained in 26 specialties.

What are themes and axes?

Themes group conversations, for example emergency situations, while axes describe the behavior a criterion grades, such as accuracy.

Details

Pricing
Free
Type
Dataset
Categories
HealthResearch
Added
Sep 16, 2026
OpenAI iconPublisherOpenAI

Built from public sources

We put this page together from HealthBench’s own website and docs so you can compare it with other tools. Ploba isn’t affiliated with HealthBench.

Work at HealthBench? Claim the listing to manage it yourself.

Something wrong here?

More from OpenAI

View publisher
ChatGPT Work iconChatGPT WorkOpenAI · FreemiumChatGPT Work is an agent in ChatGPT that takes action across your apps and files and turns a goal into finished work such as sheets, slides, docs and web apps. It can stay with a project for hours.Productivity · FreemiumOpenAI Realtime API iconOpenAI Realtime APIOpenAI · Usage-basedOpenAI's Realtime API lets developers build speech-to-speech voice agents that work directly with audio, keep conversation state and call tools. It is for developers building voice assistants and other low-latency conversational apps.Developer Tools · Usage-basedOpenAI Dots iconOpenAI DotsOpenAI · SubscriptionDots are always-on agents in ChatGPT that take on ongoing responsibilities and keep making progress between conversations. You choose which apps they use and when they need your approval.Productivity · Subscription

Related tools

See more
Our World in Data Health iconOur World in Data HealthOur World in Data · FreeOur World in Data's global health collection brings together charts, data and writing on life expectancy, child and infant mortality, and healthcare spending. Charts are open access under a Creative Commons BY license.Education · FreeAlphaMissense iconAlphaMissenseGoogle DeepMind · FreeAlphaMissense is a catalogue of predictions for human missense variants, classifying each as likely pathogenic or likely benign. It covers all 71 million possible missense variants.Health · FreePhysioNet iconPhysioNetPhysioNet · FreePhysioNet is an NIH-funded platform run by the MIT Laboratory for Computational Physiology that hosts curated, de-identified biomedical and clinical datasets, software and models. Most resources are openly available; some require credentialing.Developer Tools · Free