MLCommons Introduces Agentic Inference Benchmark for MLPerf Inference

MLCommons unveiled a new MLPerf benchmark for multi-turn LLM serving under growing context and closed-loop agent workflows.