MODEL INFO: detection of token window size and costs

This is a request that I have already made in the past without success, but I am trying again because otherwise, this information has to be manually entered into applications (damn it!).

REQUEST for windows size and cost (two options):
A) Include the two pieces of information in the data returned by “/v1/models”;
B) Use a stable model nomenclature (as has already been done for some) e.g.
GPT-X-YYYY-MM-DD-128K or
GPT-X-128K

  In this case, I understand that for commercial reasons the cost cannot be specified.

Regards.

1 Like

Welcome back! You’ve been a stranger! :wink:

I’ve switched your category to Feedback. No guarantees it will be seen by OpenAI staff, but who knows!

What have you been up to? Anything cool to share? Tips for us?

We love dev tips! :wink:

1 Like

Thank you. You are very kind. Ehm … the main things perhaps a new embedding algorithm, now in test, it seams to works well.

1 Like

Nice. That’s something!

Feel free to visit more and help others if you can. We’ve got a great dev community garden growing here, I think.

Once you hit Trust Level 3 on the forum, you get access to the Lounge, a special area for forum members where we’re a little more open in discussions and share tips, projects, etc.

Hope you stick around!

1 Like