According to DeepLearning.AI, Artificial Analysis has launched version 4.0 of its Intelligence Index, introducing new evaluation tests that focus on economically useful work, factual reliability, and reasoning. This update replaces outdated, saturated benchmarks to more accurately assess how large language models perform in real-world business scenarios. As reported by DeepLearning.AI, the new benchmarks are designed to reflect the models' capabilities in delivering value for enterprises, offering actionable insights for organizations assessing AI integration in business operations.