Developer Registration & Fees ( Token Charges )

Community
, ,
1

For GPT development is the process clear on 1. being recognized as an official “Developer”? Ive noticed they have asked for a domain website, is this required? and 2. for Token Charges, when will these charges apply? At the time of launching a GPT publicly? I want to make sure I wont have any difficulty launching my GPTs when the GPTs launch to the public.

1 Like
2
  1. Becoming an Official Developer: OpenAI’s GPT models are accessible to a broad range of users. Plus and Enterprise users can start creating GPTs via ChatGPT. There’s no explicit mention of needing to be recognized as an official “Developer.” However, for those with coding skills, there is an option for additional customization through coding actions that connect GPTs to external data or services​​​​.
  2. Token Charges and Launching GPTs: OpenAI’s pricing for GPT models is based on tokens used. A token can be thought of as a piece of a word, with 1,000 tokens being approximately 750 words. Different GPT models have different per-token costs. For example, the GPT-4 model costs $0.03 per 1,000 tokens for input and $0.06 per 1,000 tokens for output. The GPT-3.5 Turbo model, on the other hand, is cheaper, costing $0.0010 per 1,000 tokens for input and $0.0020 per 1,000 tokens for output​​​​​​​​. Additionally, there are other costs associated with specific services like the code interpreter and data retrieval, though some of these services are currently free until a specific date in 2023​​.

These charges apply as you use the services to process inputs and outputs. So, if you are planning to launch a GPT publicly, you would incur these charges as soon as your GPT starts processing data. It’s important to consider these costs in your planning for launching any GPT-based service or application.

Lastly, it’s advisable to regularly check OpenAI’s official website for the most current information, as pricing and policies can evolve over time.

1 Like
3

Thank you this was extremely helpful. When you say “as soon as your GPT starts processing data” do you mean after launch or during the construction of my GPT? Appreciate your help and feedback…

4

That I can’t answer. I can say however I have used no tokens in my testing of the GPTs functionality in the playground, so I would think that it’s a post launch topic. But don’t get me lying. I have been known to make errors.

1 Like
5

Lol. Appreciate you being so helpful. Ive been building a few GPTs but I just want to make sure Ill be able to launch them when the time comes without any issues. I havent been using the playground, only directly building GPTs. Thanks again for your feedback and help

1 Like
6

side note: the GPT builder you are currently using applies the dynamics of the playground next to the builder. When you test it as your building it its a playground test.

image
image1408×810 68.4 KB

1 Like
7

Hey, thanks for the topic. I had understood the things differently that what you mentioned.

From my understanding: GPTs are available only to the ChatGPT “Plus” users, and they would be free to use if you have a subscription. I have created some GPTs, used them like crazy but still see 0.00 as costs.

The assistants created using the “Assistant API” would be charged for usage, as these are available to everyone / public.

It would be useful to understand the prices officially… if an OpenAI staff could answer.

1 Like
8

You can edit this to not do a disservice to everyone. The AI it seems you used got it way wrong.

Only if you build your own “plugin” API for a GPT action, and then also have that API powered by some AI service that calls on paid AI every call, would it cost you much to have a GPT share or in a store.

A GPT is a “chat share” but with more shared and no chat and only for continuing Plus subscribers.

11

ChatGPT

I conducted a detailed comparison between the data you provided and the information available on OpenAI’s official website. Here’s a summary of the findings:

  1. Becoming an Official Developer:
  • Both your data and OpenAI’s website indicate that developers have the option for additional customization through coding actions. This aligns well and shows no discrepancies. [Source: OpenAI Help Center]
  1. Token Charges and Launching GPTs:
  • Your Data: Mentioned specific pricing for GPT-4 and GPT-3.5 Turbo models.
  • OpenAI’s Website: The pricing details for GPT-4 and GPT-3.5 Turbo models are slightly different. For instance, GPT-4 costs $0.03 per 1K tokens for input and $0.06 per 1K tokens for output. GPT-4 Turbo costs $0.01 per 1K tokens for input and $0.03 per 1K tokens for output. GPT-3.5 Turbo costs $0.0010 per 1K tokens for input and $0.0020 per 1K tokens for output. [Source: OpenAI Pricing]
  • Additional Services: The code interpreter and data retrieval services have specific charges mentioned on OpenAI’s website, which aligns with the information you provided. [Source: OpenAI Pricing]
  1. Data Integrity:
  • The data you received seems to be slightly outdated or incorrect in terms of specific pricing details when compared to the current information on OpenAI’s website. It’s important to note that OpenAI’s pricing and policies can evolve over time.

I recommend regularly checking OpenAI’s official website for the most current information, as pricing and policies are subject to change. Unfortunately, due to time constraints in the browsing tool, I couldn’t further verify every aspect of the GPT models in detail. However, the key points of interest regarding developer status and token charges have been addressed based on the available information.

12

Got it! Clear now. Playground is where you test your builder. Thanks for all your help on this…

1 Like
13

Its still not clear to me if I will be charged for any GPTs created based on inputs and outputs. I guess we will see at time of launch in the GPT store. The other big question is if they will allow GPT authors the ability to charge for use. I know they mentioned revenue share but that is a very vague term.

14

I watched an enlightening video on YouTube by Liam Evans. Ill leave the searching up to you but he generalized some answers and strategies. So far, I am having what I assume is a typical experience of testing and imaginative. I hadn’t been billed and only had to deal with the 50-response height ceiling which is fine. I like the playground and have some good ideas going. Feel free to dm me and I’ll show you mine if you show me yours…

1 Like
15

Sounds good. Im still working on them and will share with you once complete.