The Phi-3.5-mini-instruct (3.82 billion parameters), Phi-3.5-MoE-instruct (41.9 billion parameters), and Phi-3.5-vision-instruct (41.9 billion parameters) are the three new Phi 3.5 models. They are, respectively, intended for basic/fast reasoning, more powerful reasoning, and vision (image and video analysis) tasks. Microsoft isn't basing its success in AI on its collaboration with OpenAI.