#open source model

Microsoft introduces Phi-4-mini-Flash-Reasoning, a compact 3.8B parameter model optimized for efficient long-context reasoning and fast inference, outperforming previous models on complex benchmarks.