Documentation issues: Responses endpoint storage, storage persistence, truncation

[vb]

Well, I do read these

We recently had another topic with a similar observation. Linking it here to keep these reports together.

@WhiteRabbit

Question about context window and truncation

Reading the official documentation on reasoning models, I found the following diagram: [context-window] (Source: https://platform.openai.com/docs/guides/reasoning?api-mode=responses#how-reasoning-works) As far as I know, the context window in LLM works in such a way that when generating each next token, the last N tokens are taken, but no more than the window size. That is, the window seems to slide constantly during generation, taking only the last tokens within the limit. But then it turns…