Question: How can I switch among the gpt 5.5 models?

There is not a separate model name required.

“Long context” is simply an increased amount of billing per token when you send an input over the previous 272k that models before gpt-5.4 had as an input limit.

Now gpt-5.4 and gpt-5.5 can accept up to 1M context length input, with an automatic switch in pricing when you pass the previous input threshold.

You do not need to pick a special model; you only need to send more input to have your bill for the entire API call increased by 100%/50%.

Reminder, not only do you have a batch input file size limit of 200 MB, but your API organization will also have a token queue limit for the batch endpoint - and the usage of API limits after passing the 272k threshold of input on calls is also doubled, or when indicated a separate model in your limits, halved.