Increase API Response Limit

[hasty hearth]

Hi guys.

With the help of ChatGPP I have written a text that describes my requirements for some code, functions and so on - which is quite a bit.
In the playground he wants to write me file by file. Yesterday I had not yet added a payment method and at some point I came to the 8000 token limit.
If I then remove code in the playground, which he has already written, the rest of the newly written code is not based on the same code and he wants to write files partially again.
Is there a way to increase the tokens for a preset? That he writes out all the files - if it costs me something, that's also ok. The response does not even have to be so long (token wise, though it would be preferable), only if I press "submit" that he gradually writes all files to the end.
I would like to use the preset then of course also for debugging and changes. I will compress my descriptive text a bit later, but I think I might still get to the token limit like yesterday, if that would not change as a paid user?

[amber pasture]

It does not change. The token limit is a limitation of the model itself.

You can rethink your strategy. Maybe generating one file at a time?

[hasty hearth]

Yes, if I do that they don't match together as the AI don't know the code it wrote already

Isn't there a way to pay for a increasing for code-davinci-002?

[hasty hearth]

I mean, I tried that, removing code from the playground it already wrote and have the AI generate next files, but resulting in:
The AI wants to partially write files again and new ones doesn't match code for the ones written already

I wouldn't mind paying good to have token response increased on a preset for me

[amber pasture]

It’s just not possible with current models.

[amber pasture]

I’m just spitballing here, but perhaps sending some kind of context of previous generations to new generations

Not a trivial problem, but you can try to make it work

[hasty hearth]

It's something modular, OOP based
So I guess it all need to match together

Is it planned to update models in the near future to support that, if you know that? Or do you have any ideas/suggestions for me?
I mean it's not possible to have files sendet directly so I need the model to write it for me

[amber pasture]

Surely the models will support more tokens as techniques and scale progresses

[hasty hearth]

I'm pretty new to AI's and don't understand how the models work with the tokens yet completely, so I'm not sure for what the limit is required etc

[hasty hearth]

Current technologies shouldn't be a problem
And would pay for the scale, hope they will come up with a solution soon 😄