MLCommons is a nonprofit artificial intelligence (AI) engineering consortium that brings together companies, universities, and researchers to collaboratively improve the performance, safety, and reliability of AI systems. Founded in 2020, MLCommons grew out of the MLPerf benchmarking initiative, which began in 2018 to create standardized ways of measuring machine learning performance.
The organization operates as a 501(c)(6) nonprofit consortium built on a philosophy of open collaboration. Through collective engineering efforts, MLCommons develops shared benchmarks, datasets, and tools that enable researchers and organizations to measure and improve the accuracy, safety, speed, and efficiency of AI technologies.
A central focus of MLCommons is the development of MLPerf, a widely used benchmarking suite that evaluates machine learning systems across a range of tasks and computing environments. These benchmarks allow organizations to compare AI system performance using transparent and standardized metrics, helping advance the maturity and reliability of machine learning technologies.
MLCommons works internationally, with more than 100 member organizations spanning industry, academia, and research institutions. Through collaborative working groups, members contribute to the development of benchmarks, open datasets, and best practices that support responsible and effective AI development.
By providing shared measurement frameworks and open engineering resources, MLCommons works to strengthen the global AI ecosystem and supports the development of more reliable and trustworthy machine learning systems.