How to Get Gemini to Write Long-Form Content?
Learn how to make Google Gemini generate 10,000+ word articles instead of short responses. Complete guide with API settings and prompts.
"AI Quill" Publication 200 Subscriptions 20% Discount Offer Link.
"Why has my Gemini output become so short? Compared to 0325, it seems to output much fewer words at once, often just two to three thousand tokens, whereas before it could reach over ten thousand tokens."
This is a common issue. Honestly, I don't frequently need ultra-long outputs in my daily use - I'm more accustomed to brief outputs as they allow me to guide AI corrections at any time.
However, I must admit that when writing novels, sometimes we do need to output complete long texts of tens of thousands of words at once, so that in key descriptive passages, we have greater space for expansion and trimming.
Gemini's output length "personality" actually fluctuates frequently across different environments. This doesn't mean the official team has cut the model's output capabilities.
In reality, a model's basic capability parameters usually don't change easily, but when server computing power and the model's internal structure change after iterations, the overall "personality" shifts. Yet our configuration methods as users remain unchanged, so when these two factors combine, they inevitably produce vastly different results.
Given this, users can only continuously fine-tune their input prompts to adapt as much as possible to the model's new personality.
So, how should we configure settings to make Gemini rise again?