AIJune 17, 2025OpenBMB Releases MiniCPM4: Ultra-Efficient Language Models for Edge Devices with Sparse Attention and Fast Inference
