MLPerf Client 2.0: AI PC Benchmarking for Image Gen & Agentic AI

MLCommons has released MLPerf Client version 2.0, an updated benchmark for evaluating AI performance on personal computers. This new version expands AI PC benchmarking to include image generation and agentic AI. The benchmark measures how effectively PCs run AI workloads locally. Version 2.0 introduces a new image generation category and a new agentic AI category. It also updates existing LLM inference tests, including an intermediate summarization task with a 4K token input prompt. These updates aim to provide a more comprehensive, cross-platform benchmark for client AI computing. This matters because it helps consumers understand and compare the AI capabilities of different PCs.