OpenAI’s GPT-4 might be coming to an end. Here’s why that’s actually good news

OpenAI has seen many changes in recent weeks, and more are quickly coming. The AI company has yet to confirm the launch of its upcoming GPT-5 AI model. However, it is making room for its planned model by ending support for other models in its lineup. OpenAI recently announced that it is retiring its GPT-4 AI model as of April 30. GPT-4 stood as one of the brand’s most popular and longest-running large language models. However, the company has already shifted its focus away from its original large language model technology and more toward its series of reasoning models and other technologies in recent months. 

The brand has also made some interesting moves by introducing a new GPT 4.1 model family, strictly as an API for developers, while simultaneously indicating plans to sunset the recently launched GPT-4.5 model and also releasing the o3 and o4 reasoning models. While not yet confirmed, these moves appear to propel the GPT-5 timeline closer to launch.

GPT-4 History 

OpenAI released GPT-4 in March 2023 as the model behind the ChatGPT Plus paid subscription option in the early days of the AI revolution. The company introduced ChatGPT Plus amid high demand for the chatbot, which caused strain on its servers. During this time, ChatGPT Plus subscribers got access to GPT-4, while free ChatGPT users got access to the original GPT-3.5 model that OpenAI launched during its initial introduction of the AI service.

While OpenAI noted several improvements between GPT-3.5 and GPT-4, including an increased context window from 2048 tokens to up to 128k tokens, some of the highlight upgrades to GPT-4 included the ability to upload images for analysis via text answers, and more creative text results after inputting more detailed prompts. There are several version updates to the GPT-4 model, including GPT-4 Turbo and GPT-4 Turbo with Vision, as well as GPT-4o and GPT-4o mini. GPT-4o mini is currently the base model across all ChatGPT tiers.

Introduction of reasoning models 

OpenAI began an interesting journey in introducing its reasoning models in December 2024, during its 12 Days of OpenAI marketing campaign. While the event unveiled several new developments for the brand, rolling out its o1 reasoning model was a standout moment. The introduction was similar to the public launch of its large language models, giving consumers a rundown of the new form of AI. Reasoning models are better at logical thinking and can show the thought process behind AI results. In contrast, large language models are trained on data to provide a contextual response to a query. During the 12 Days event, OpenAI also introduced its $200 per month ChatGPT Pro tier and made the o1 model exclusive to that pricing option. 

Since then, OpenAI has announced its new o3 and o4-mini reasoning models with improved performance in coding, math, and science tasks, in comparison to the o1 model. They are also available within the $20 ChatGPT Plus tier, and the o4-mini model is included within the free tier of ChatGPT. 

GPT-4.5 and the Turing test 

While OpenAI has shown an affinity for reasoning models, it has also continued to develop and release its LLMs. The brand made a statement by launching GPT-4.5 in late February 2025 as its largest AI model to date. The technology was notable for its decreased hallucinations, a persistent issue for AI models. It also displayed increased conversation and problem-solving capabilities, as well as a greater emotional prowess. It has the ability to pick up on nuances in creativity in tasks such as writing and design. The model was available to ChatGPT Plus and ChatGPT Pro users.

In early April 2025, reports indicated that GPT-4.5 passed the Turing Test intelligence evaluation in a UC San Diego study. The test entails human participants selecting which model they thought was human and which they thought was AI after five minutes of engagement. Researchers uncovered that when set to mimic the profile of a “geeky 19-year-old knowledgeable about internet culture,” the AI model was selected as human 73% of the time. Without a profile, the model was selected as human 36% of the time.  

The Turing test is considered important because it allows researchers to study the potential applications and ethical ramifications of AI in real time. The study determined that through the GPT-4.5 model, humans could be able to have successful conversations with AI. 

However, OpenAI announced its decision to retire GPT-4.5 in mid-April, just months after making the model available. The company said it didn’t meet the “frontier level” of AI that it was aiming for, with technology exceeding current models. OpenAI plans to fully remove GPT-4.5 from its model lineup, having advised developers that they will have access to the API until July 14, and giving guidance to transition their projects to other APIs. 

API efforts 

While GPT-4 may be retired soon, OpenAI has detailed that its API will still be available for developers. Currently, it remains unknown how the brand will use the protocol in the future. However, it may be an easy option for building consumer-focused AI tools. Duolingo, Stripe, and Microsoft have already used GPT-4 to power AI features on their products. OpenAI also recently made its GPT‑3.5 Turbo API available to developers. This is the base model that powered the ChatGPT chatbot when it was first introduced to the public in 2022. Similarly, brands including Snap Inc., Quizlet, Instacart, and Shopify have used the API to make consumer-facing AI features. 

OpenAI has also recently announced GPT-4.1 in the API, a developer-focused model family, which includes GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano, and offers the same functionalities as GPT-4.5 at a lower cost. OpenAI claims it has performance updates over GPT-4o; however, it has no major improvements to be called a “frontier level.” Even so, the model is knowledgeable up to June 2024. OpenAI has collaborated with brands including Windsurf, Qodo, Hex, Blue J, Thomson Reuters, and Carlyle to test GPT-4.1’s capabilities as its fastest and cheapest model. 

GPT-5 plans 

GPT-5 has been highly anticipated for some time, and many of the efforts OpenAI has put into its previous models may culminate in significant improvements for the upcoming AI model. The company hasn’t detailed a specific timeline for its launch; however, sometime in 2025 looks more plausible than ever. The brand indicated early in the year that GPT-5 could be available in the coming months. 

Prior reports suggested the AI model could arrive sometime in May or during the summer. That seems unclear now as the company has released its o3 and o4-mini reasoning models. However, OpenAI is known for its surprises. The company’s CEO, Sam Altman, has even hinted on X that there might be some updates to AI model naming sequences before GPT-5 releases.

There are indications that OpenAI is working to have GPT-5 be a frontier level model, bringing something new and unique to the industry. Under previous parameters, GPT-4 is considered a frontier AI model. The reasoning models also have that title, shifting the capacity of how AI can process data.

Altman has indicated that GPT-5 will have autonomous AI agents among its highlight features that will aid users by executing tasks independently. In another anticipated feature, GPT-5 may have the ability to select the best available model based on your query. 

Users on the free tier of ChatGPT are also expected to have limited access to the GPT-5 model. However, those with Plus and Pro subscriptions will really be able to take advantage of the coming features.