Difference of GPT's and Assistants

can i understand it like GPTs are with a Frontend and Assistants more for the Backend/Internal ? any specific use-cases?

  • GPT: for the new plugin store in ChatGPT Plus
  • Assistants: for API developers who can’t code want to write more code.

As Assistants can be used on API, how can I integrate a custom GPT to my website? And also, Why are GPTS much more precise in their responses than the assistant when I upload documents?


Agree - this is the question I have as well. What is the relationship between GPTs and Assistants & Tools. I see GPT’s consuming Assistants & Tools while the GPT is focused more on the flow and tone. But hey, I’m happy to be corrected.

Likewise, I want to know if my GPT has an API endpoint. I would like to be able to call from UI in other apps that I build.


I’d say it depends really on your use case and the complexity it involves. I have been working with both GPTs and Assistant APIs this week. GPTs are great and really straightforward to create (although I am still figuring out the actions part) but there are of course certain limitations to it. This is where Assistant APIs come in really handy as they offer more tools and flexibility.

awesome meeting you here. i think so too. GPTs more for the final user, Assistant more for the coders.

i built an aws ec2, ubuntu, nginx, python, php, mongodb chat app.
it could do multimodel before openai made its public, was fun, and my latest feature ’ public chat bot’ now also from openai via gpt’s. that makes me feel even more motivated to work on my own one. what to build to keep different for a longer time, unique, valueable? :slight_smile:

i love the digital journey.
riding it since over 40 years.

what is this:
beeppppp tratratrtrat beeepp beeepppppppp tratttarrata zzzz…

Thanks @ilovesinai and @jr.2509 - here is my situation - I am teaching a class and I want each student to submit their final paper as a GPT that is infused with their knowledge from the class - they are building an Assistant that does something “generative”. I want them to be able to plug their Assistant into their GPT - so the GPT functions as a sort of interactive documentation of their assistant and also allow the GPT user to use the functionality of the Assistant. Does that make sense? I don’t want them to have to recreate the wheel to embed the functionality into the GPT… Thanks in advance

1 Like

i think i get you. basically you give them the task to create a chatbot, that is connected to other assistent apis or whatever and does extra tasks, is that right?
recreating the wheel is harder than to keep going from where we are.
my own app is doing this already without gpt’s and assistants from openai, source code build 99% by ai itself. such things make me excited.

as i saw u can call functions from the gpt’s so u could call a function from there to a function of the assistant. . or what exactly is the question :slight_smile:

This sounds fun!

Currently GPTs and the API Assistants are not interchangeable, even though they share the same functionality (on the surface)

I believe there will be “group” ChatGPT accounts which is similar to an enterprise account and allows multiple people to use it. This may be the best option.

Otherwise you’ll need to find/create a GUI for the API along with the backend system (user identification & database)

1 Like

That’s what I do with my ai app. building a userbase that can exchange information and data.

It looks like GPTs will become OpenAI new plugins, and evolve to a free and paid apps market. The user will pay to openai, and a revenue share model with the developer will exist.

Assistants do the same, but accessible with the API so they can be integrated into your own app. You will pay OpenAI for usage. Your customers will pay you.

What’s best is dependent on exact use case.


I would add that GPTs are restricted to access only for ChatGPT+ users - so you limit a lot the range of people who will have access to it - whereas Assistants enable everyone to play with it through the Playground, as well as enable you to create interfaces on top of it to be shared with anyone through any digital channel.

1 Like

I think that GPTs are more likely to Plugins, and assistants are mroe likely for dev use

1 Like

Absolutely thrilled to be part of this dynamic group!

So, what’s everyone diving into?
I’m curious about your AI endeavors and the goals you’re aiming to achieve.

And let’s not forget the hurdles – what are those for you?

Actual I am crafting a pioneering AI based app, with the aim to make a catalyst for exchanging value and needs – a platform where you can store and share your offerings, either privately or with the world. Think about a simple marketplace for values, needs, and ideas, where the NLP of the AI is ur guide.

Currently in the exciting beta phase, with the prototype exceeding expectations in functionality; while also working on some known bugs. lol.

Among the features, partly available already, is a shared database for notes, reminders, and specialized chat bots, designed to streamline the experience.

My vision? To create a standalone, community-centric AI app. A platform that not only serves humanity but also seamlessly integrates with existing systems, forming a network of value and innovation. https://gotomy.ai is my sandbox around the corner should u be interested to pass by.

happy sunday :tada::smiling_face_with_three_hearts::peace_symbol::pray:

Do you do work for hire? :slight_smile: I am trying to integrate assistant with my joomla website and cannot get a response when I publish the module on the website.

1 Like

I encountered the same problem. With the same model, parameters and uploaded the same file, the quality of response from GPTs is much better than assistant. I don’t why

1 Like

Do you have hourly rates or would you meet with me online and let me show you the issue so you can figure a fee? I have a meet server if needed.

To followup on the question: How would the two differ from the Pricing standpoint upon public usage? For instance, are GPTs running for free to start with while AI assistant usage to be billed to developers?

Yes, GPTs are free (minus your subscription).

The downside is that GPTs are restricted to the ChatGPT interface and the user must also have a plus subscription.

1 Like

Could someone please explain to me if MemGPT is orthogonal to using the Assistants API or if they can meaningfully be used together or if MemGPT is in some way superseded by these new announcements?