IBM has released two new language models, PowerLM-3B and PowerMoE-3B, featuring 3 billion parameters and an advanced power scheduler designed for efficient large-scale AI training. These models represent a significant leap in efforts to improve the efficiency of language models. Concurrently, OpenBMB has introduced MiniCPM3-4B, a versatile and efficient small language model with advanced functionality, extended context handling, code generation capabilities, better mathematical ability, and proficiency. MiniCPM3-4B also offers scalability in model and data dimensions. The release of these models highlights the ongoing advancements in AI and machine learning, emphasizing the importance of both large and small language models in the current technological landscape.