Microsoft just released a new 15-billion parameter open-weight AI model that can do something most bigger models can't.
Google introduces Gemini 3.1 Flash-Lite in preview via AI Studio and Vertex AI, promising faster responses and lower costs for high-volume apps.