Newshot

InclusionMed, the World's First Independent Evaluation and Analysis Platform for Healthcare AI, Officially Launches

AI summary

InclusionMed, the world's first independent evaluation and analysis platform for healthcare AI, has officially launched.

Translation(model translation, not the source's own text)

Article title: World's First Independent Medical and Health AI Evaluation and Analysis Platform InclusionMed Officially Launched

Body:

World's First Independent Medical and Health AI Evaluation and Analysis Platform InclusionMed Officially Launched

Recently, InclusionMed, the world's first independent evaluation and analysis platform for medical and health AI, was officially launched

Recently, InclusionMed, the world's first independent evaluation and analysis platform for medical and health AI, was officially launched, and simultaneously released the medical AI evaluation map InclusionMed Atlas and the first medical AI evaluation index InclusionMed Index, aiming to provide a more systematic evaluation basis for medical and health AI. The platform was jointly built by researchers from 29 institutions including Stanford University, Harvard University, Peking Union Medical College, Ant Group, and InclusionMed.AI, and is regarded as the Artificial Analysis (AA) of the medical and health field, filling the gap in comprehensive evaluation of global medical and health AI.

Can an AI that performs well in health Q&A also do complex professional medical tasks equally well? How can it identify key clues in difficult cases, accurately complete operations in electronic medical record systems, and handle multi-step tasks in medical research and healthcare operations? Each scenario places different demands on models. For model capabilities under specific medical tasks, there are already various evaluation benchmarks in the industry, but single-test scores are difficult to fully reflect a model's performance across different medical tasks.

It is understood that InclusionMed Atlas, centered on the two dimensions of "for whom" and "what capabilities are examined," crosses three roles—patients and the public, medical professionals, and healthcare administrators—with three types of capabilities—knowledge, reasoning, and task execution—forming a nine-grid evaluation map. Each benchmark is assigned to the corresponding position according to its main evaluation content, helping users see which medical needs current evaluations cover and which aspects still need supplementation. InclusionMed Index, meanwhile, aggregates the evaluation results of the included benchmarks, presenting the model's overall performance and its strengths and weaknesses across different tasks.

The first evaluation focuses on the two major directions of clinical diagnosis and treatment and healthcare management, includes five evaluation benchmarks and 1,034 tasks, systematically compares 22 models from 11 manufacturers at home and abroad, and after independent re-testing, comprehensively examines factors such as medical relevance, evaluation credibility, and scenario complementarity, ultimately selecting four public evaluation benchmarks, which together with the team's self-built InclusionMed-DxFiles constitute the first index. Among them, InclusionMed-DxFiles is built based on difficult cases published in top medical journals, examining the model's ability to make clinical judgments by combining case clues.

It is understood that the project team is collaborating with medical institutions and research teams and plans to develop new evaluation tasks around scenarios such as doctor-patient shared decision-making in complex treatment choices and diagnosis and treatment scheduling under resource constraints, to further examine AI's performance in clinical diagnosis and treatment, patient decision support, and healthcare management.

This article is provided by Ant, and QbitAI is authorized to reprint it; the views belong to the original author.

QbitAI's friends
QbitAI's friends
  • Realman Intelligence 郑随兵: Quality, economy, and stability are the three major standards for robot deployment 2026-11-16
  • The 2026 Jiazi Gravity Annual Gala is about to open, and AI's infinite game enters the "Second Prologue" | Jiazi Gravity 2026-11-16
  • SuperLink2026 opens the venture capital industry's super connection moment, see you in Beijing in December 2026-11-16
  • Alibaba officially releases Qwen4, with advanced agent and coding capabilities, moving toward real all-scenario applications 2026-11-16

Related Reading

Ant Code Large Model Commercialization: Fully Integrated with SOFAStack, Institutional R&D Efficiency Increased by 30%

End-to-end Copilot product solution

This Double 11, Bailing's New Model Gets to Work Without Burning Out

Whether a large model is easy to use, Bailing lets the finished product speak

Ant Digital Technologies CTO Wang Wei: Trusted Agents Have Become a Necessary Infrastructure for Scaled Industrial Application of AI

The underlying logic of enterprise service products is about to undergo fundamental change.

Ant Invests in an Embodied Intelligence Company That Makes Hands

It had just received the industry's largest seed round of funding four months ago

The All-People AI Health Steward Is Here! Hands-On Test of Ant AQ: Follow-Up Questions to Identify Medicines and Examine Skin, Even Connecting Hospitals and Hardware

200 famous doctors' avatars provide uninterrupted consultations all day

Replicating AI IMO Gold Medal Results in 6 Hours, Ant's New Multi-Agent Progress Has Been Open-Sourced

IMO2025 Revelation: Will Multi-Agent "Swarm Intelligence" Surpass Single Models?

The translation is machine-generated and individual terms may be off. Nothing is rewritten or cut; the original is authoritative.

Published by 量子位
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗