【#Tech24H】Google has officially launched the Gemma 4 family of large language models, claimed to be its most intelligent opensource models to date. Gemma 4 is designed for advanced reasoning and agentic workflows, delivering an unprecedented level of “intelligence per parameter”. Google introduces four generalpurpose Gemma 4 variants: E2B, E4B, MoE (26B Mixture of Experts), and 31B (31B dense model). Among these, the 26B and 31B models are tailored for highend servers equipped with powerful GPUs (e.g., NVIDIA H100), while the E2B and E4B models are primarily aimed at mobile and IoT devices. Their smaller parameter footprint minimizes impact on device performance, enabling efficient operation on mobile and edge devices.