The End of ChatGPT Plugins: Will GPTs Be Sufficient Replacements?

Goodbye ChatGPT plugins, hello OpenAI’s GPTs

ChatGPT plugins have been replaced by generative pre-trained transformers (GPTs), custom versions of ChatGPT that anyone can create for specific purposes as no coding is required. As of March 19, 2024, users can no longer create new conversations with OpenAI plugins; however, users can continue using existing plugin conversations until April 9, 2024.

“With the launch of GPTs and the GPT Store, we were able to make many improvements that plugin users had asked for,” according to OpenAI, developer of the ChatGPT chatbot. 

“GPTs now have full feature parity (in addition to many new features) with plugins.”

OpenAI did not respond to a request for comment from Techopedia about replacing ChatGPT plugins with GPTs.

Key Takeaways

  • ChatGPT plugins have been replaced by GPTs, custom versions of ChatGPT that anyone can create for specific purposes.
  • GPTs are large language models that are specifically trained to generate human-like content.
  • With the launch of GPTs and the GPT Store, OpenAI was able to make many improvements that plugin users had asked for.
  • The GPT Store already has more than 3 million custom versions of ChatGPT in a wide variety of categories, such as productivity, programming, writing, education, and more.

Advertisements

ChatGPT Plugins vs. GPTS

ChatGPT plugins were additional functionalities or integrations built on top of OpenAI’s ChatGPT large language model

They extended ChatGPTs capabilities, offering such features as conversation management, content generation, language translation, task automation, sentiment analysis, personalization, knowledge retrieval, and security enhancements. 

These plugins were developed by platform providers, third-party developers, or users to tailor ChatGPT to specific needs or use cases.

Now, OpenAI has developed a set of large language models (LLMs) known as GPTs that are specifically trained to comprehend and produce human-like content that’s based on the input they are given. GPTs use transformer architecture to effectively process and generate text while being aware of the context. 

These LLMs have been pre-trained on massive quantities of internet text data. Users can adjust them for specific tasks and applications. Individuals and organizations can use GPTs for a diverse range of natural language processing tasks, including translating language, generating text, analyzing sentiment, summarizing, and more.

During the ChatGPT plugins beta, there were slightly more than 1,000 plugins available in the ChatGPT plugin store, according to OpenAI. However, the GPT Store already has more than 3 million custom versions of ChatGPT in a wide variety of categories, such as productivity, programming, writing, education, and more. 

“Since launching ChatGPT [Nov. 30, 2022], people have been asking for ways to customize ChatGPT to fit specific ways that they use it,” according to OpenAI. “We launched custom Instructions in July [2023] that let you set some preferences, but requests for more control kept coming. Many power users maintain a list of carefully crafted prompts and instruction sets, manually copying them into ChatGPT. GPTs now do all of that for you.”

Developer Marcelo Arias says If ChatGPT plugins haven’t been converted to GPTs, they will be removed. Arias is a software engineer at SmartUp, a startup that helps other companies integrate OpenAI technologies into the communication flow with their customers.

“The conversion is quite simple for a software developer,” he says. “So if they don’t convert them, it’s probably because they don’t want to.” 

Arias told Techopedia: 

“I still offer advice on how GPTs can benefit other parts of a business to companies that contact me. Thanks to the fact that I was one of the first developers on the list, and one of the first to publish their plugins, I was able to find a lot of growth in my network of contacts.”

Are Custom GPTs Better Than ChatGPT Plugins?

As ChatGPT plugins transform into advanced generative process transformers, AI capabilities will be transformed, says Mark Gadala-Maria, co-founder and CEO of Post Cheetah, an AI-enabled SEO tool. 

“GPTs offer better language understanding, promising contextually accurate and precise text generation,” he says. “This represents a significant shift toward more powerful language models, as evidenced by the adoption of the GPT-3.5 and GPT-4 models.”

As plugin alpha access and customization options increase, GPTs are designed to adapt and evolve to meet diverse business needs, according to Gadala-Maria. He says:

“As AI continues to grow, GPTs are becoming more and more powerful, enabling them to fully replace and even increase the capabilities of ChatGPT plugins.”

ChatGPT plugins made it easier to focus the language models on specific problems, such as math, or synonyms, or browsing, says R “Ray” Wang, principal analyst and founder of Constellation Research Inc.

“[However], you could only run three at once and it was limiting, but it was a great way to get started. Going forward, you won’t have to be restricted and it’s all in the GPT.”

Users Weigh In

After learning that OpenAI was phasing out ChatGPT plugins in favor of GPTs, a number of users offered their opinions in the OpenAI developer forum.

For example, in January, user Richard22 wrote:

Right now I use several plugins daily in the same context; one to read PDFs, one to search the web, one to analyze GitHub, and a couple of different ones to draw diagrams, also Wolfram Alpha sometimes (obviously there is a maximum of [three] per chat, I mix and match depending on the task).

He adds that while none of these is particularly useful on its own, combined they are “amazing.

“Am I correct in saying that GPTs will have no way of replicating this functionality of using a variety of plugins together in the same context?” Richard22 asks.

While OpenAI did not respond to this question, user Elmstedt did respond.

“You make an interesting observation,” Elmstedt notes. “With plugins users were able to mix-and-match functionality provided by different developers – this was very powerful and very useful. It was also unpredictable and uncontrollable.”

Elmstedt adds that It’s certainly possible that some combination of plugins could result in some kind of unaligned behavior/outputs from the models.

“I’m not suggesting this is why plugins are being phased out, just making an observation,” he says.

And according to user Sreuter, “[GPTs] pretty much allow for everything you could do via plugins (via actions) plus allow you to provide more context, connect things with browsing the web, set conversation status. Plus, they are much more discoverable than plugins.”

Sreuter adds that custom GPTs provide a much better experience over plugins, given that enabling and disabling plugins inside a conversation all the time felt really cumbersome.

Techopedia reached out to the user, Sreuter, who is Sascha Reuter, founder of Questmate, an app for managing and sharing recurring tasks. Reuter pointed us to a blog post he authored on this subject.

“Fundamentally, custom GPTs come with pre-configured actions and capabilities, e.g., do something in app ‘X,’ as well as specific embedded knowledge for the task they have been built for,” he notes. “That also means that instead of users having to turn plugins on and off manually, they can switch conversations depending on the task, which usually is way quicker and more intuitive.”

Five Main Reasons GPTs Are Better Than Plugins

In his blog post, Reuter offers five main reasons GPTs are significantly better than plugins for users and developers.

  1. Open to Everyone

    “The biggest thing that a lot of developers waiting to get invited by OpenAI to write traditional plugins are missing is that you can not only do the same but more with GPTs,” he says. “But also in contrast to publishing plugins, GPTs can be listed on the ‘GPT Store’ or simply by sharing a direct URL, which ultimately results in way more GPTs being made available than plugins.”
  2. Knowledge

    “Instead of cluttering your system message (or instructions), adding knowledge to your GPT is as easy as uploading a file,” Reuter says.
  3. Capabilities

    Depending on the context of GPT, Reuter says users can enable or disable specific in-built capabilities, such as web browsing, DALL-E image generation, and code interpreter, from ChatGPT for the conversations.
  4. Conversation Starters

    In the past, users had to try plugins so they could understand how to use them. Now, however, Reuter says that by using easy-to-add conversation starters, GPT creators can give users a couple of suggestions based on what their GPTs are good at.
  5. No Code, No Servers

    “For best results, you’d still want to integrate with your product API, especially if you aim to get more users to discover your product and services via the GPT,” Reuter says. However, he adds that for pure knowledge-driven GPTs, a creator has to add an “instruction” paragraph (to provide the GPT with necessary context) and then upload their knowledge as files.

The Bottom Line

“Many users are getting used to GPTs and are finding a lot more value in them,” Wang says. “Some users still like the simplicity and utility of focused plugins. However, over time ChatGPT will be successful in replacing the plugins.”

Software developer Arias says that GPTs are definitely a greater abstraction than plugins: 

“In a way, plugins have become the actions component of every GPT.”

FAQs

What is the difference between GPT and ChatGPT plugins?

Are plugins still used?

Advertisements

Related Reading

Related Terms

Advertisements
Linda Rosencrance
Technology journalist

Linda Rosencrance is a freelance writer and editor based in the Boston area, with expertise ranging from AI and machine learning to cybersecurity and DevOps. She has been covering IT topics since 1999 as an investigative reporter working for several newspapers in the Boston metro area. Before joining Techopedia in 2022, her articles have appeared in TechTarget, MSDynamicsworld.com, TechBeacon, IoT World Today, Computerworld, CIO magazine, and many other publications. She also writes white papers, case studies, ebooks, and blog posts for many corporate clients, interviewing key players, including CIOs, CISOs, and other C-suite execs.