I am trying to use the gemini-2.5-flash-lite model, but the API returns the following error:
{"error":{"code":404,"message":"This model models/gemini-2.5-flash-lite is no longer available to new users. Please update your code to use a newer model for the latest features and improvements.","status":"NOT_FOUND"}}
What I find confusing is that Gemini’s own documentation still describes this model as:
“The fastest and most budget-friendly multimodal model in the 2.5 family.”
Being advertised as the most budget-friendly option in the 2.5 family is precisely what makes this model suitable for my use case, where request cost and scalability are important factors.
I would like to understand why gemini-2.5-flash-lite is no longer available to new users while it remains documented and, apparently, still available to existing users.
Is this restriction temporary? Is there an eligibility requirement? Is the model being gradually deprecated?
I would also like to understand why new users are being directed toward newer and significantly more expensive models while a lower-cost option is still documented.
From the perspective of someone building a commercial product on top of the Gemini API, this raises a legitimate concern: lower-cost models may become unavailable to new projects while more expensive alternatives remain accessible.
I do not think a generic explanation based solely on infrastructure capacity would adequately address the question, considering the scale of Google’s infrastructure.
An official explanation of this decision and the model availability policy would be very helpful, particularly regarding whether developers can expect models advertised as “budget-friendly” to remain available to new projects in the future.