Published November 1, 2026 | Version v1

MatProGPT database

Authors/Creators

Description

This dataset contains structured battery material property data extracted from the scientific literature using MatProGPT, a multi-module large language model framework designed to retrieve material properties from scientific papers with high accuracy. The pipeline was applied to extract more than 524,926 structured data points from 47,000 articles.

Files

MatProGPT_dataset_all_properties.csv

Files (27.3 MB)

Name Size
md5:6e20f823313122137f74be6514e7ebd7
27.3 MB Preview Download

Additional details

Dates

Available
2027-04-01