New AI benchmarks test speed of running AI applications

As the underlying models must respond to many more queries to power AI applications such as chatbots and search engines, MLCommons developed two new versions of its MLPerf benchmarks to gauge speed