OpenAI is pulling older ChatGPT models following GPT-5 launch - so bad news if you use GPT-4 or others at work

News
By published

Enterprise customers have a week left to say goodbye to GPT-4

GPT-5
(Image credit: OpenAI)
  • OpenAI says GPT-5 is rolling out to all users, replacing previous models with immediate effect
  • ChatGPT now includes automatic model routing via GPT-5 models
  • Enterprise APIs still include access to older models - but not for long

Following OpenAI's launch of GPT-5, its most advanced model to date, the AI giant has removed a number of previous-generation models from its repertoire with immediate effect.

The launch of GPT-5 has seen OpenAI officially retire its previous-generation models (GPT-4o, o3, o4-mini and o4-mini-high) for many users, instead moving to make GPT-5 the standard option across all tiers including free accounts.

However, it now means users will no longer be able to pick from previous-generation models for their AI tools, which they may have preferred for specific workloads.

GPT-5 replaces GPT-4o and other previous models

The launch comes as ChatGPT approaches 700 million weekly users, including five million ChatGPT paying business product users, whose lives it promises to revolutionize by unlocking "greater productivity, efficiency, and creative output."

With the launch of OpenAI's latest model, users will no longer be able to pick specific models like GPT-4o for certain tasks, with GPT-5 instead promising to be a unified agent that can respond quickly and reason deeply, all in one place.

Enterprise and Edu users will retain access to legacy models for 60 days, with GPT-5 access only arriving a week after its initial launch, meaning they'll only have a short time to consider how to continue moving forwards.

Enterprises using the API will also retain access to previous-generation models, which are not planned to be deprecated for now.

According to OpenAI's announcement, GPT-5 can now outperform many workers across around 40 fields, including law, sales, logistics and engineering, with around 45% fewer hallucinations than GPT-4o.

A number of iterations are also launching, including different preset personalities and tones within GPT-5, a deeper GPT-5 Thinking and GPT-5 Pro, which is available to Pro tier subscribers. GPT-5 mini will also serve as a fallback for lower-tier and free users who meet their regular GPT-5 limit.

Even though the GPT-5 family promises a broad range of applications and improved performance, many enterprise users say GPT-4o's performance was best for casual tasks – a model they will no longer be able to use.

Surely, though, the change was due to come, as OpenAI CEO Sam Altman had previously criticized the model picker for being too complex, hinting at unification.

As models improve, including with the launch of GPT-5, OpenAI envisions a world where businesses can realize "better decision-making, improved collaboration, and faster outcomes on high-stakes work."

You might also like

Craig Hale
Craig Hale

With several years’ experience freelancing in tech and automotive circles, Craig’s specific interests lie in technology that is designed to better our lives, including AI and ML, productivity aids, and smart fitness. He is also passionate about cars and the decarbonisation of personal transportation. As an avid bargain-hunter, you can be sure that any deal Craig finds is top value!

You must confirm your public display name before commenting

Please logout and then login again, you will then be prompted to enter your display name.