Uh.. does anyone else notice reasoning changes in Gemini 2.5 pro? Either they switched the model this morning or they are hiding the chain of thought (It's now displaying summaries versus specifics, a la OpenAI models)