datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
OpenMedallion
OpenMedallion Financial Dataset
Comprehensive financial dataset for quantitative research, machine learning, and trading strategy backtesting.
2,609 Parquet files · 18 categories · 1,500+ unique assets · 87+ countries · Up to 100 years of history
Quick Start
import pandas as pd
# Load BTC 1h data (5 years)
df = pd.read_parquet("data/BTC-BTCUSD_1h.parquet")
# Load Indonesian stocks
df = pd.read_parquet("data/equities/country_stocks/BBCA_1d.parquet")
# Load Gold… See the full description on the dataset page: https://huggingface.co/datasets/oyi77/OpenMedallion.Medical-Reasoning-SFT-Mega
Medical-Reasoning-SFT-Mega
The ultimate medical reasoning dataset - combining 7 state-of-the-art AI models with fair distribution deduplication. 1.79 million unique samples with 3.78 billion tokens of medical chain-of-thought reasoning.
Dataset Overview
Metric
Value
Total Samples
1,789,998 (after deduplication)
Total Tokens
~3.78 Billion
Content Tokens
~2.22 Billion
Reasoning Tokens
~1.56 Billion
Samples with Reasoning
1,789,764 (100.0%)
Unique… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-Mega.synthvision-annotated-qwen
synthvision-annotated-qwen
Medical images annotated by Qwen 3.5 (397B) via Doubleword
Records: 59,476
About
First-half annotations from the SynthVision pipeline. 59,476 medical images annotated by Qwen 3.5 (397B MoE, 17B active) via Doubleword batch inference.
Each record contains a multi-turn clinical conversation (5-9 turns), a clinical narrative report, structured findings, reasoning chain, and difficulty rating.
Schema
id: str #… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/synthvision-annotated-qwen.synthvision-validated-qwen-by-kimi
synthvision-validated-qwen-by-kimi
Qwen 3.5 annotations validated by Kimi K2.5 (93.1% pass rate)
Records: 55,359
About
Cross-validated subset from the SynthVision pipeline. Kimi K2.5 reviewed all 59,476 Qwen 3.5 annotations and confirmed 55,359 as consistent with the source images (93.1% pass rate).
Validation criteria: consistent == true AND confidence >= 0.7. Records that failed validation were removed — primarily cases where the annotator hallucinated findings not… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/synthvision-validated-qwen-by-kimi.Medical-Reasoning-SFT-GPT-OSS-120B
Medical-Reasoning-SFT-GPT-OSS-120B
A high-quality synthetic dataset of medical reasoning conversations generated using OpenAI's gpt-oss-120B model with reasoning effort set to high, designed for supervised fine-tuning of large language models in healthcare applications. I used Intelligent-Internet/II-Medical-Reasoning-SFT as a seed dataset, so I would like to thank the authors and Intelligent-Internet for their great work.
Dataset Statistics
Total Samples: 200,927… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-GPT-OSS-120B.OpenMedReason
This data is the Open-PMC derived data part that we discuss in the paper
Medical VQA with Reasoning Traces (Anonymous)
A multiple-choice medical visual question answering benchmark with
chain-of-thought reasoning traces. Each example consists of a medical image
(radiology, pathology, clinical photograph, etc.), a multiple-choice question
with labeled options, a reasoning trace, and the correct answer letter.
This dataset is released anonymously in support of a… See the full description on the dataset page: https://huggingface.co/datasets/researcher2026/OpenMedReason.MedDialog
MedDialog
A large-scale medical dialogue dataset containing ~252k patient-doctor conversation pairs for training and evaluating clinical dialogue systems.
Dataset Description
Property
Value
Source
ruslanmv/ai-medical-chatbot
License
Apache-2.0
Language
English
Total examples
251,731
Train split
226,557
Validation split
25,174
Domain
Clinical / General Medicine
Overview
MedDialog is designed for training language models to generate… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/MedDialog.synthvision-validated-kimi-by-qwen
synthvision-validated-kimi-by-qwen
Kimi K2.5 annotations validated by Qwen 3.5 (93.0% pass rate)
Records: 55,382
About
Cross-validated subset from the SynthVision pipeline. Qwen 3.5 reviewed all 59,539 Kimi K2.5 annotations and confirmed 55,382 as consistent with the source images (93.0% pass rate).
Validation criteria: consistent == true AND confidence >= 0.7. Records that failed validation were removed — primarily cases where the annotator hallucinated findings not… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/synthvision-validated-kimi-by-qwen.synthvision-seeds
synthvision-seeds
Seed records from 4 open medical image datasets
Records: 119,137
About
Seed dataset for the SynthVision pipeline. Contains 119,137 records aggregated from 4 open medical image datasets:
Source
Records
Modality
eltorio/ROCO-radiology
65,393
Radiology
OpenMed/multicare-images
50,000
Mixed
flaviagiammarino/path-vqa
3,430
Pathology
flaviagiammarino/vqa-rad314
Radiology
Images are deduplicated by SHA-256 hash. Each record contains an… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/synthvision-seeds.synthvision-annotated-kimi
synthvision-annotated-kimi
Medical images annotated by Kimi K2.5 via Doubleword
Records: 59,539
About
Second-half annotations from the SynthVision pipeline. 59,539 medical images annotated by Kimi K2.5 (1T MoE, 32B active) via Doubleword batch inference.
Each record contains a multi-turn clinical conversation (5-9 turns), a clinical narrative report, structured findings, reasoning chain, and difficulty rating.
Schema
id: str # unique… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/synthvision-annotated-kimi.Medical-Reasoning-SFT-Nemotron-Nano-30B
Medical-Reasoning-SFT-Nemotron-Nano-30B
A large-scale medical reasoning dataset generated using nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16, containing over 444,000 samples with detailed chain-of-thought reasoning for medical and healthcare questions.
Dataset Overview
Metric
Value
Model
nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
Total Samples
444,544
Samples with Reasoning
444,544 (100%)
Estimated Tokens
~1.01 Billion
Content Tokens
~808 Million… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-Nemotron-Nano-30B.Medical-Reasoning-SFT-GPT-OSS-120B-Small
Medical-Reasoning-SFT-GPT-OSS-120B-Small
A filtered and processed version of OpenMed/Medical-Reasoning-SFT-GPT-OSS-120B optimized for training efficiency.
Dataset Description
This dataset contains high-quality medical reasoning conversations with the following modifications:
Length Filtering: Only includes samples where assistant responses are between 1000 and 10000 characters
Reasoning Extraction: Reasoning content from <think> tags has been extracted into a separate… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-GPT-OSS-120B-Small.openmed-traces
OpenMed audit traces
This private dataset stores opt-in OpenMed Agent audit traces owned by this
Hugging Face account or organization.
Default enabled traces store full private transcript and tool payloads unless
the local OpenMed audit config is explicitly set to preview, fingerprint, or
none.
Raw audit runs are written under runs/YYYY/MM/DD/<run_id>.json.
Viewer-facing root JSONL files are stable per session and are updated as the
session grows. schema.json describes the… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/openmed-traces.TheBlueScrubs-v1-fixed
openmed-community/TheBlueScrubs-v1-fixed
What is this?
TheBlueScrubs-v1-fixed is a maintenance fork of the upstream TheBlueScrubs/TheBlueScrubs-v1 train split that resolves a schema bug in the meta column.In the original train files, some rows serialized meta incorrectly (appearing as the literal string "dict"). This fork re-exports the entire train split without meta column, preserving text field and values.
Document count: 11,080,331 texts (train)
Tokens (upstream… See the full description on the dataset page: https://huggingface.co/datasets/openmed-community/TheBlueScrubs-v1-fixed.Medical-Reasoning-SFT-Trinity-Mini
Medical-Reasoning-SFT-Trinity-Mini
A large-scale medical reasoning dataset generated using arcee-ai/Trinity-Mini, containing over 810,000 samples with detailed chain-of-thought reasoning for medical and healthcare questions.
Dataset Overview
Metric
Value
Model
arcee-ai/Trinity-Mini
Total Samples
~810,374
Estimated Tokens
~1.52 Billion
Content Tokens
~542 Million
Reasoning Tokens
~977 Million
Language
English
Schema
Each… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-Trinity-Mini.MedReason-Stenographic
MedReason-Stenographic: Medical QA with Compressed Reasoning Traces
This dataset contains 31,535 medical question-answer pairs with stenographic reasoning traces, generated using MiniMax M2.1 from the original UCSC-VLAA/MedReason dataset.
Dataset Description
The dataset transforms medical QA reasoning into a stenographic format using a symbolic protocol designed for high-density, machine-parseable reasoning traces. This format eliminates natural language filler while… See the full description on the dataset page: https://huggingface.co/datasets/openmed-community/MedReason-Stenographic.multicare-images
MultiCaRe: Open-Source Clinical Case Dataset
MultiCaRe is an open-source, multimodal clinical case dataset built from the PubMed Central Open Access (OA) Case Report articles. It aggregates de-identified, open-access case narratives, figure images, captions, and rich article metadata across diverse specialties (radiology, pathology, surgery, ophthalmology, etc.). The data is normalized so images, cases, and articles can be joined via stable IDs.
Source and process: OA case reports… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/multicare-images.Medical-Reasoning-SFT-GPT-OSS-120B-V2
Medical-Reasoning-SFT-GPT-OSS-120B-V2
A large-scale medical reasoning dataset generated using openai/gpt-oss-120b, containing over 506,000 samples with detailed chain-of-thought reasoning for medical and healthcare questions.
GPT-OSS-120B is OpenAI's state-of-the-art open-weight model, achieving near-parity with closed models on reasoning benchmarks while being Apache 2.0 licensed.
Dataset Overview
Metric
Value
Model
openai/gpt-oss-120b
Total Samples
506… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-GPT-OSS-120B-V2.Medical-Reasoning-SFT-Qwen3-Next-80B
Medical-Reasoning-SFT-Qwen3-Next-80B
A large-scale medical reasoning dataset generated using Qwen/Qwen3-Next-80B-A3B-Thinking, containing over 604,000 samples with detailed chain-of-thought reasoning for medical and healthcare questions.
Dataset Overview
Metric
Value
Model
Qwen/Qwen3-Next-80B-A3B-Thinking
Total Samples
604,249
Samples with Reasoning
604,249 (100%)
Estimated Tokens
~1.42 Billion
Content Tokens
~505 Million
Reasoning Tokens
~917 Million… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-Qwen3-Next-80B.Medical-Reasoning-SFT-Baichuan-M3-235B
Medical-Reasoning-SFT-Baichuan-M3-235B
A large-scale medical reasoning dataset generated using baichuan-inc/Baichuan-M3-235B, containing over 124,000 samples with detailed chain-of-thought reasoning for medical and healthcare questions.
Baichuan-M3-235B is ranked #1 on HealthBench Total leaderboard and achieves state-of-the-art performance on medical reasoning benchmarks.
Dataset Overview
Metric
Value
Model
baichuan-inc/Baichuan-M3-235B
Total Samples
124… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-Baichuan-M3-235B.medical-raw
Medical QA Dataset
Overview
This dataset is a collection of multiple Medical QA sources, benchmarks, mock tests, and extracted data from various PDFs. It is intended for research and development in medical question-answering tasks.
⚠ Important Note: Only the MedQA and MedMCQA datasets have been filtered to exclude their respective test sets. Other sources may still contain test data, so use caution when evaluating models.
Dataset Sources
The dataset is… See the full description on the dataset page: https://huggingface.co/datasets/OpenMedical/medical-raw.OpenMedReason
OpenMedReason
Scientific Reasoning Supervision for Medical Vision-Language Models
Medical visual question answering with structured, image-grounded chain-of-thought. Each
example pairs a biomedical image with a multiple-choice question and a reference reasoning
trace that moves from visual evidence → medical knowledge → clinical reasoning → answer.
The held-out test split adds a per-example unit-question rubric to measure how a
model… See the full description on the dataset page: https://huggingface.co/datasets/neginb/OpenMedReason.medical-verifiable-dedup
Medical QA Dataset
Overview
This dataset is a collection of multiple Medical QA sources, benchmarks, mock tests, and extracted data from various PDFs. It is intended for research and development in medical question-answering tasks.
We deduplicated based on UUID from text using MD5.
⚠ Important Note: Only the MedQA and MedMCQA datasets have been filtered to exclude their respective test sets. Other sources may still contain test data, so use caution when evaluating… See the full description on the dataset page: https://huggingface.co/datasets/OpenMedical/medical-verifiable-dedup.m1-raw-medical-reasoning
m1-raw QA Dataset
Overview
This is a synthetic dataset generated using state-of-the-art (SOTA) models (<32B) on our medical QA dataset. Each question has three generated answers.
We used Sglang to generate the data, leveraging code from the open-r1 project by Hugging Face.
⚠ Important Note: This dataset is unfiltered and may contain problematic content. Please apply filtering before use.
Usage Considerations
The dataset is intended for research… See the full description on the dataset page: https://huggingface.co/datasets/OpenMedical/m1-raw-medical-reasoning.drugprot-parquet
DrugProt (Parquet)
A clean, ready-to-use Parquet version of the DrugProt corpus from BioCreative VII Track 1, converted for seamless use with the Hugging Face datasets library.
DrugProt is a gold-standard corpus of 4,250 PubMed abstracts annotated for drug/chemical–protein interactions, covering 13 fine-grained relation types and 3 entity types. It is designed for training and evaluating biomedical relation extraction systems.
Quick Start
from datasets import… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/drugprot-parquet.agab-db
AgAb DB: Antigen Specific Antibody Database
A comprehensive collection of antibody-antigen interaction data for computational biology and therapeutic design.
Dataset Summary
AgAb DB aggregates antibody-antigen binding data from multiple sources, containing over 1.2 million antibody-antigen pairs with binding affinity measurements. This dataset is essential for training machine learning models in computational immunology and antibody engineering.
Key Statistics… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/agab-db.PubHealth-Processedm1-stage1-medical-reasoning
Analysis
Medical-Reasoning-SFT-MiniMax-M2.1
Medical-Reasoning-SFT-MiniMax-M2.1
A large-scale medical reasoning dataset generated using MiniMaxAI/MiniMax-M2.1, containing over 204,000 samples with detailed chain-of-thought reasoning for medical and healthcare questions.
Dataset Overview
Metric
Value
Model
MiniMaxAI/MiniMax-M2.1
Total Samples
204,773
Samples with Reasoning
204,773 (100%)
Estimated Tokens
~621 Million
Content Tokens
~344 Million
Reasoning Tokens
~277 Million
Language
English… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/Medical-Reasoning-SFT-MiniMax-M2.1.multicare-cases
MultiCaRe: Open-Source Clinical Case Dataset
MultiCaRe is an open-source, multimodal clinical case dataset derived from PubMed Central’s Open Access (OA) Case Report articles. It links de-identified case narratives to figure images/captions and article-level metadata, enabling cross-modal supervision and retrieval.
Source and process: OA case reports from PMC; parsed metadata and abstracts; extracted case narratives; downloaded and processed figures; aligned captions; curated image… See the full description on the dataset page: https://huggingface.co/datasets/OpenMed/multicare-cases.