Why Does Google’s Own Model Treat “NanoBanana” Like a Mystery Keyword?

Whenever I write a prompt that includes the word “NanoBanana,” Gemini’s thinking process shows a whole section where it starts “investigating” what NanoBanana is, as if it’s an unknown term: it starts trying to determine whether NanoBanana is real software, a typo, etc., instead of just treating it as it's own model name.

I’m not an AI expert, so maybe this is just how its internal reasoning or chain-of-thought works. But I haven’t really seen similar behavior on other models like Grok, ChatGPT, or Claude.. when I mention their own model names.

Is this just Gemini being extra literal, or does Google’s own model really not recognize the “NanoBanana” as referring to itself? Curious what other think.