On April 2, 2026, Google rolled out the Gemma 4 open-source AI model. This model boasts four distinct specifications, operates under the permissive Apache 2.0 license, delivers markedly enhanced performance, and is well-suited for a multitude of application scenarios.