Awesome-AI4Med

KO | EN

gitlite — search

#react #typescript #go #machine-learning #rust #deep-learning #docker #nodejs #java #windows #kubernetes #vue #pytorch #ios

(★ 2,868)

A curated list of medical LLMs, multimodal systems, datasets, benchmarks, and more. 🏥

File Explorer

Download Latest Version (.zip)

- English_vision.md
- README.md
- pathology_datasets.jpg
- pretrain_overview.png
README.md

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Awesome-AI4Med

> We have renamed *Medical_NLP* to *Awesome-AI4Med*. If you need to view the previous content, you can visit [backup](https://github.com/FreedomIntelligence/Medical_NLP/tree/master/backup)!

> [!TIP]
> **Awesome-AI4Med** aims to systematically curate research resources in the field of medical artificial intelligence (AI4Med), covering **🧠 Medical LLMs**, **🩺 Medical MLLMs**, **📚 datasets**, and **📊 benchmarks**. By automatically extracting trending models and datasets from a large corpus of research papers and open-source projects, this repository provides a **clean, structured, and easy-to-navigate** collection that helps researchers and developers **🚀 quickly track field progress**, **🔍 locate relevant resources**, and **accelerate innovation** in medical AI.
>
> 📣 You can view our repository on specialized disease LLMs at [FreedomIntelligence/Awesome-Specialized-Medical-LLMs](https://github.com/FreedomIntelligence/Awesome-Specialized-Medical-LLMs)!

Table of Contents:
- [1. Medical LLMs](#1-medical-llms)
- [2. Medical MLLMs](#2-medical-mllms)
- [3. Datasets](#3-datasets)
  - [3.1 Text](#31-text)
  - [3.2 Multimodal](#32-multimodal)
- [4. Benchmarks](#4-benchmarks)

![](https://camo.githubusercontent.com/2722992d519a722218f896d5f5231d49f337aaff4514e78bd59ac935334e916a/68747470733a2f2f692e696d6775722e636f6d2f77617856496d762e706e67)

## 1. Medical LLMs

|**Model Name**|**Paper**|**Model Weight**|**Data**|**Code**|
|:-|:-:|:-:|:-:|:-:|
|`HuatuoGPT`|[[arXiv](https://arxiv.org/abs/2305.15075)]|[[HF (7B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT-7B)][[HF (13B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT-13b-delta)]|[[HF](https://huggingface.co/datasets/FreedomIntelligence/HuatuoGPT-sft-data-v1)]|[[Github](https://github.com/FreedomIntelligence/HuatuoGPT)]|
|`Apollo`|[[arXiv](https://arxiv.org/abs/2403.03640)]|[[HF (7B)](https://huggingface.co/FreedomIntelligence/Apollo-7B)][[HF (34B)](https://huggingface.co/FreedomIntelligence/Apollo-34B)][[HF (72B)](https://huggingface.co/FreedomIntelligence/Apollo-72B)]|[[HF](https://huggingface.co/datasets/FreedomIntelligence/ApolloCorpus)]|[[Github](https://github.com/FreedomIntelligence/Apollo)]|
|`HuatuoGPT-II`|[[arXiv](https://arxiv.org/abs/2311.09774)]|[[HF (7B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT2-7B)][[HF (13B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT2-13B)][[HF (34B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT2-34B)]|[[HF (PT)](https://huggingface.co/datasets/FreedomIntelligence/HuatuoGPT2-Pretraining-Instruction)][[HF (SFT)](https://huggingface.co/datasets/FreedomIntelligence/HuatuoGPT2-SFT-GPT4-140K)]|[[Github](https://github.com/FreedomIntelligence/HuatuoGPT-II)]|
|`HuatuoGPT-o1`|[[arXiv](https://arxiv.org/abs/2412.18925)]|[[HF (7B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT-o1-7B)][[HF (8B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT-o1-8B)][[HF (70B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT-o1-70B)]|[[HF](https://huggingface.co/datasets/FreedomIntelligence/medical-o1-reasoning-SFT)]|[[Github](https://github.com/FreedomIntelligence/HuatuoGPT-o1)]|
|`MMedLM`|[[arXiv](https://arxiv.org/abs/2402.13963)]|[[HF (8B)](https://huggingface.co/Henrychur/MMed-Llama-3-8B)][[HF (70B)](https://huggingface.co/Henrychur/MMed-Llama3.1-70B)]|[[HF](https://huggingface.co/datasets/Henrychur/MMedC)]|[[Github](https://github.com/MAGIC-AI4Med/MMedLM)]|
|`PMC-LLaMA`|[[arXiv](https://arxiv.org/abs/2304.14454)]|[[HF (13B)](https://huggingface.co/axiong/PMC_LLaMA_13B)]|[[HF](https://github.com/allenai/s2orc)]|[[Github](https://github.com/chaoyi-wu/PMC-LLaMA)]|
|`Med42-v2`|[[arXiv](https://arxiv.org/abs/2408.06142)]|[[HF (8B)](https://huggingface.co/m42-health/Llama3-Med42-8B)][[HF (70B)](https://huggingface.co/m42-health/Llama3-Med42-70B)]|-|[[Github](https://github.com/m42-health/med42)]|
|`OpenBioLLM`|-|[[HF (8B)](https://huggingface.co/aaditya/Llama3-OpenBioLLM-8B)][[HF (70B)](https://huggingface.co/aaditya/Llama3-OpenBioLLM-70B)]|-|-|
|`UltraMedical`|[[arXiv](https://arxiv.org/abs/2406.03949)]|[[HF (8B)](https://huggingface.co/TsinghuaC3I/Llama-3-8B-UltraMedical)][[HF (70B)](https://huggingface.co/TsinghuaC3I/Llama-3-70B-UltraMedical)]|[[HF](https://huggingface.co/datasets/TsinghuaC3I/UltraMedical)]|[[Github](https://github.com/TsinghuaC3I/UltraMedical)]|
|`BioMistral`|[[arXiv](https://arxiv.org/abs/2402.10373)]|[[HF (7B)](https://huggingface.co/BioMistral/BioMistral-7B)]|-|[[Github](https://github.com/BioMistral/BioMistral)]|
|`MediTron`|[[arXiv](https://arxiv.org/abs/2311.16079)]|[[HF (7B)](https://huggingface.co/epfl-llm/meditron-7b)][[HF (70B)](https://huggingface.co/epfl-llm/meditron-70b)]|[[HF](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T)]|[[Github](https://github.com/epfLLM/meditron)]|
|`Baichuan-M2`|[[arXiv](https://arxiv.org/abs/2509.02208)]|[[HF (32B)](https://huggingface.co/baichuan-inc/Baichuan-M2-32B)]|-|[[Github](https://github.com/baichuan-inc/Baichuan-M2-32B)]|
|`Baichuan-M1`|[[arXiv](https://arxiv.org/abs/2502.12671)]|[[HF (14B)](https://huggingface.co/baichuan-inc/Baichuan-M1-14B-Instruct)]|-|[[Github](https://github.com/baichuan-inc/Baichuan-M1-14B)]|
|`MedAlpaca`|[[arXiv](https://arxiv.org/abs/2304.08247)]|[[HF (7B)](https://huggingface.co/medalpaca/medalpaca-7b)][[HF (13B)](https://huggingface.co/medalpaca/medalpaca-13b)]|[[Github](https://github.com/kbressem/medAlpaca/blob/main/DATA_DESCRIPTION.md)]|[[Github](https://github.com/kbressem/medAlpaca)]|
|`RadFM`|[[arXiv](https://arxiv.org/abs/2308.02463)]|[[HF (13B)](https://huggingface.co/chaoyi-wu/RadFM)]|[[Github](https://huggingface.co/datasets/chaoyi-wu/RadFM_data_csv)]|[[Github](https://github.com/chaoyi-wu/RadFM)]|
|`MedReason`|[[arXiv](https://arxiv.org/abs/2504.00993)]|[[HF (8B)](https://huggingface.co/UCSC-VLAA/MedReason-8B)]|[[HF](https://huggingface.co/datasets/UCSC-VLAA/MedReason)]|[[Github](https://github.com/UCSC-VLAA/MedReason)]|
|`AlphaMed`|[[arXiv](https://www.arxiv.org/abs/2505.17952)]|[[HF (7B)](https://huggingface.co/che111/AlphaMed-7B-instruct-rl)][[HF (8B)](https://huggingface.co/che111/AlphaMed-8B-instruct-rl)]|[[HF](https://huggingface.co/datasets/che111/AlphaMed19K)]|-|
|`Fleming-R1`|[[arXiv](https://arxiv.org/abs/2509.15279)]|[[HF (7B)](https://huggingface.co/UbiquantAI/Fleming-R1-7B)][[HF (32B)](https://huggingface.co/UbiquantAI/Fleming-R1-32B)]|-|[[Github](https://github.com/UbiquantAI/Fleming-R1)]|
|`MedGemma`|[[arXiv](https://arxiv.org/abs/2507.05201)]|[[HF (27B)](https://huggingface.co/google/medgemma-27b-text-it)]|-|[[Github](https://github.com/Google-Health/medgemma)]|
|`AntAngelMed`|-|[[HF (100B)](https://huggingface.co/MedAIBase/AntAngelMed)]|-|[[Github](https://github.com/MedAIBase/AntAngelMed)]|
|`Baichuan-M3`|-|[[HF (235B)](https://huggingface.co/baichuan-inc/Baichuan-M3-235B)]|-|-|
|`MedPsy`|[Blog](https://huggingface.co/blog/qvac/medpsy)|[[HF (1.7B/4B)](https://huggingface.co/collections/qvac/medpsy)]|-|-|
|`MediPhi`|[[arXiv](https://arxiv.org/abs/2505.10717)]|[[HF (4B)](https://huggingface.co/collections/microsoft/mediphi)]|-|-|

## 2. Medical MLLMs

|**Model Name**|**Paper**|**Model Weight**|**Data**|**Code**|
|:-|:-:|:-:|:-:|:-:|
|`HuatuoGPT-Vision`|[[arXiv](https://arxiv.org/abs/2406.19280)]|[[HF (7B)](https://huggingface.co/FreedomIntelligence/HuatuoGPT-Vision-7B)][[HF 34B](https://huggingface.co/FreedomIntelligence/HuatuoGPT-Vision-34B)]|[[HF](https://huggingface.co/datasets/FreedomIntelligence/PubMedVision)]|[[Github](https://github.com/FreedomIntelligence/HuatuoGPT-Vision)]|
|`ShizhenGPT`|[[arXiv](https://arxiv.org/abs/2508.14706)]|[[HF (7B)](https://huggingface.co/FreedomIntelligence/ShizhenGPT-7B-Omni)][[HF (32B)](https://huggingface.co/FreedomIntelligence/ShizhenGPT-32B-VL)]|[[HF (PT)](https://huggingface.co/datasets/FreedomIntelligence/TCM-Pretrain-Data-ShizhenGPT)][[HF (SFT)](https://huggingface.co/datasets/FreedomIntelligence/TCM-Instruction-Tuning-ShizhenGPT)]|[[Github](https://github.com/FreedomIntelligence/ShizhenGPT)]|
|`Lingshu`|[[arXiv](https://arxiv.org/abs/2506.07044)]|[[HF (7B)](https://huggingface.co/lingshu-medical-mllm/Lingshu-7B)][[HF 32B](https://huggingface.co/lingshu-medical-mllm/Lingshu-32B)]|-|-|
|`LLaVA-Med`|[[arXiv](https://arxiv.org/abs/2306.00890)]|[[HF (7B)](https://huggingface.co/microsoft/llava-med-v1.5-mistral-7b)]|[[Github](https://github.com/microsoft/LLaVA-Med/tree/main/data)]|[[Github](https://github.com/microsoft/LLaVA-Med)]|
|`BiMediX2`|[[arXiv](https://arxiv.org/abs/2412.07769v2)]|[[HF (8B)](https://huggingface.co/MBZUAI/BiMediX2-8B)][[HF (70B)](https://huggingface.co/MBZUAI/BiMediX2-70B)]|[[Github](https://github.com/microsoft/LLaVA-Med/tree/main/data)]|[[Github](https://github.com/mbzuai-oryx/BiMediX2)]|
|`Hulu-Med`|[[arXiv](https://arxiv.org/abs/2510.08668)]|[[HF (4B)](https://huggingface.co/ZJU-AI4H/Hulu-Med-4B)][[HF (7B)](https://huggingface.co/ZJU-AI4H/Hulu-Med-7B)][[HF (14B)](https://huggingface.co/ZJU-AI4H/Hulu-Med-14B)][[HF (32B)](https://huggingface.co/ZJU-AI4H/Hulu-Med-32B)]||[[Github](https://github.com/ZJUI-AI4H/Hulu-Med)]|
|`HealthGPT`|[[arXiv](https://arxiv.org/abs/2412.07769v2)]|[[HF (3.8B)](https://huggingface.co/lintw/HealthGPT-M3)]|[[HF](https://huggingface.co/datasets/lintw/VL-Health)]|[[Github](https://github.com/DCDmllm/HealthGPT)]|
|`Med-Flamingo`|[[arXiv](https://arxiv.org/abs/2307.15189)]|[[HF (7B)](https://huggingface.co/med-flamingo/med-flamingo)]|-|[[Github](https://github.com/snap-stanford/med-flamingo)]|
|`MedGemma`|[[arXiv](https://arxiv.org/abs/2507.05201)]|[[HF (4B)](https://huggingface.co/google/medgemma-4b-it)][[HF (27B)](https://huggingface.co/google/medgemma-27b-it)]|-|[[Github](https://github.com/Google-Health/medgemma)]|
|`Med-R1`|[[arXiv](https://arxiv.org/abs/2503.13939)]|[[HF (2B)](https://huggingface.co/yuxianglai117/Med-R1)]|-|[[Github](https://github.com/Yuxiang-Lai117/Med-R1)]|
|`BiomedGPT`|[[arXiv](https://arxiv.org/abs/2305.17100)]|[[HF (7B)](https://huggingface.co/PharMolix/BioMedGPT-LM-7B)]|[[Github](https://github.com/taokz/BiomedGPT/blob/main/datasets.md)]|[[Github](https://github.com/taokz/BiomedGPT)]|
|`MedVLM-R1`|[[arXiv](https://arxiv.org/abs/2502.19634)]|[[HF (2B)](https://huggingface.co/JZPeterPan/MedVLM-R1)]|-|[[Github](https://github.com/JZPeterPan/MedVLM-R1)]|
|`UniMedVL`|[[arXiv](https://arxiv.org/abs/2510.15710)]|[[HF (14B)](https://huggingface.co/General-Medical-AI/UniMedVL)]|-|[[Github](https://github.com/uni-medical/UniMedVL)]|
|`Chiron-o1`|[[arXiv](https://arxiv.org/abs/2506.16962)]|[[HF (2B)](https://huggingface.co/manglu3935/Chiron-o1-2B)][[HF (8B)](https://huggingface.co/manglu3935/Chiron-o1-8B)]|-|[[Github](https://github.com/manglu097/Chiron-o1)]|
|`MiniGPT-Med`|[[arXiv](https://arxiv.org/abs/2407.04106)]|[[HF](https://github.com/Vision-CAIR/MiniGPT-Med?tab=readme-ov-file#download-minigpt-med-trained-model-weights)]|[[Github](https://github.com/Vision-CAIR/MiniGPT-Med?tab=readme-ov-file#dataset)]|[[Github](https://github.com/Vision-CAIR/MiniGPT-Med)]|
|`Fleming-VL`|[[arXiv](https://arxiv.org/abs/2511.00916)]|[[HF (8B)](https://huggingface.co/UbiquantAI/Fleming-VL-8B)][[HF (38B)](https://huggingface.co/UbiquantAI/Fleming-VL-38B)]|-|[[Github](https://github.com/UbiquantAI/Fleming-VL)]|
|`VILA-M3`|[[arXiv](https://arxiv.org/abs/2411.12915)]|[[HF (3B)](https://huggingface.co/MONAI/Llama3-VILA-M3-3B)][[HF (8B)](https://huggingface.co/MONAI/Llama3-VILA-M3-8B)][[HF (13B)](https://huggingface.co/MONAI/Llama3-VILA-M3-13B)]|-|[[Github](https://github.com/Project-MONAI/VLM-Radiology-Agent-Framework)]|
|`MedGemma-1.5`|[[Blog](https://research.google/blog/next-generation-medical-image-interpretation-with-medgemma-15-and-medical-speech-to-text-with-medasr/)]|[[HF (4B)](https://huggingface.co/google/medgemma-1.5-4b-it)]|-|[[Github](https://github.com/google-health/medgemma)]|
|`MediX-R1`|[[arXiv](https://arxiv.org/abs/2602.23363)]|[[HF (2B)](https://huggingface.co/MBZUAI/MediX-R1-2B)][[HF (8B)](https://huggingface.co/MBZUAI/MediX-R1-8B)][[HF (30B)](https://huggingface.co/MBZUAI/MediX-R1-30B)]|[[HF](https://huggingface.co/datasets/MBZUAI/medix-rl-data)]|[[Github](https://github.com/mbzuai-oryx/MediX-R1)]|
|`MedMO`|[[arXiv](https://arxiv.org/abs/2602.06965)]|[[HF (4B)](https://huggingface.co/MBZUAI/MedMO-4B)][[HF (8B)](https://huggingface.co/MBZUAI/MedMO-8B)][[HF (8B-Next)](https://huggingface.co/MBZUAI/MedMO-8B-Next)]|-|[[Github](https://github.com/genmilab/MedMO)]|
|`MedGPT-oss`|[[arXiv](https://arxiv.org/abs/2603.00842)]|-|-|-|
|`OctoMed`|[[arXiv](https://arxiv.org/abs/2511.23269)]|[[HF (7B)](https://huggingface.co/OctoMed/OctoMed-7B)]|-|-|
|`QoQ-Med`|[[arXiv](https://arxiv.org/abs/2506.00711)]|[[HF (7B)](https://huggingface.co/ddvd233/QoQ-Med-VL-7B)][[HF (32B)](https://huggingface.co/ddvd233/QoQ-Med-VL-32B)]|-|[[Github](https://github.com/DDVD233/QoQ_Med)]|
|`MedVLThinker`|[[arXiv](https://arxiv.org/abs/2508.02669)]|[[HF (3B/7B/32B)](https://huggingface.co/collections/UCSC-VLAA/medvlthinker)]|[[HF](https://huggingface.co/collections/UCSC-VLAA/medvlthinker-688f52224fb7ff7d965d581d)]|[[Github](https://github.com/UCSC-VLAA/MedVLThinker)]|
|`Citrus-V`|[[arXiv](https://arxiv.org/pdf/2509.19090)]|[[HF (8B)](https://huggingface.co/jdh-algo/Citrus-V-8B-v1.0)]|[[HF](https://huggingface.co/datasets/jdh-algo/MeCoVQA-G-Plus)]|[[Github](https://github.com/jdh-algo/Citrus-V)]|
|`ClinFusion`|[[arXiv](https://arxiv.org/abs/2607.24743)]|[[HF (8B)](https://huggingface.co/Alibaba-DAMO-Academy/ClinFusion-8B)][[HF (32B)](https://huggingface.co/Alibaba-DAMO-Academy/ClinFusion-32B)]|-|[[Github](https://github.com/alibaba-damo-academy/ClinFusion)]|

## 3. Datasets

### 3.1 Text

### 3.2 Multimodal

## 4. Benchmarks

|**Evaluation Framework**|**Code**|**Supported Benchmark**|
|:-|:-:|:-:|
|`MedEvalKit`|[[Github](https://github.com/alibaba-damo-academy/MedEvalKit)]|**（1）Multimodal：**：`MMMU-Medical-test`, `MMMU-Medical-val`, `PMC_VQA`, `OmniMedVQA`, `IU XRAY`, `MedXpertQA-Multimodal`, `CheXpert Plus`, `MIMIC-CXR`, `VQA-RAD`, `SLAKE`, `PATH-VQA`, `MedFrameQA`；**（2）Text-only：**：`MedQA-USMLE`, `MedMCQA`, `PubMedQA`, `Medbullets-op4`, `Medbullets-op5`, `MedXpertQA-Text`, `SuperGPQA`, `HealthBench`, `CMB`, `CMExam`, `CMMLU`, `MedQA-MCMLE`；|
|`MedBench`|[[Website](https://medbench.opencompass.org.cn/home)]|**Text-only：**：`Med-Exam`, `MedHC`, `MedMC`, `MedSpeQA`, `MedHG`, `MedDG`, `MedLitQA`, `IMCS-V2-MRG`, `DBMHG`, `CMB-Clin-extended`, `DDx-basic`, `DDx-advanced`, `MedTreat`, `CMeEE`, `CMeIE`, `CHIP-CDEE`, `CHIP-CDN`, `CHIP-CTC`, `SMDoc`, `MedSafety`, `DrugCA`, `CriID`|
|`LiveClin`|[[Github](https://github.com/AQ-MedAI/LiveClin)]|-|
|`Stanford MedHELM⁠`|[[Website](https://crfm.stanford.edu/helm/medhelm/latest/#/leaderboard)]|Holistic Evaluation of Language Models (HELM)|
|`Medmarks v0.1⁠`|[[Website](https://sophont.med/blog/medmarks#introduction)]|-|
|`HealthBench⁠`|[[Github](https://github.com/medicalsphere/HealthBench)]|`healthbench`, `healthbench_hard`, `healthbench_consensus`, `healthbench_meta`, `healthbench_professional`|

Awesome-AI4Med

File Explorer

Today's picks

Was this content helpful?

Recent Feedback

Awesome-AI4Med

File Explorer

Today's picks

Was this content helpful?

Recent Feedback

Download README