When AI answers suddenly get short and perfunctory, do not rush to switch models: most of the time the cause is one of three things — settings, context, or quota. Work through them in order and you can usually pinpoint the cause within five minutes.
First: check whether conversation settings changed
Check whether the current conversation has a terse-answer mode on, or a system instruction like “answer as briefly as possible.” Some tools reset preferences after an update. Start a fresh conversation and put the requirement into the question, for example “please expand in detail and list the reasons point by point.” If the new conversation behaves normally, the problem was in the old conversation's settings.
Second: check whether the context has grown too long
The longer a conversation runs, the tighter the usable context becomes, and answers get shorter and thinner. Copy out the key conclusions and continue in a new conversation, or ask the model to summarize the discussion so far before continuing. For long-document Q&A, split the document into parts and ask about each separately; that works better than dumping the whole text at once.
Third: check quota and model version
An exhausted free quota, a downgrade to a lighter model, or peak-hour throttling can all shrink answers noticeably. Check your account quota, the current model name, and network status, then try a different time slot or switch back to the flagship model. If all three checks are fine and switching models does not help, consider whether the question itself is too vague: state the background, the goal, and the desired output format clearly, so the model has something to expand on.
State your requirements in the question
Many short answers are not laziness: the question itself left no room to expand. Try this structure: one sentence of background, one sentence of goal, plus format and length, for example asking for a plain-language, point-by-point explanation of a concept with examples and misconceptions, around 800 words. Putting requirements in the question works far better than following up with “be more detailed,” which often just produces longer empty talk.
