OpenAI’s GPT-4 model is set to be “fully replaced” by the natively multimodal GPT-4o model. According to ChatGPT’s release notes, GPT-4 will be phased out from ChatGPT on April 30, although it will remain accessible through the API. Recent enhancements to GPT-4o have made GPT-4 relatively obsolete.
The release notes highlight that updates have significantly improved GPT-4o’s abilities in instruction following, problem-solving, and conversational flow, positioning it as a natural successor to GPT-4.
In the context of the rapidly evolving AI sector, the transition underscores the pace at which the industry is advancing, with GPT-5 still pending release. GPT-4, introduced in March 2023, marked a significant progression from GPT-3.5 and facilitated the widespread debut of ChatGPT. Since then, OpenAI has introduced the multimodal GPT-4o and various “o-series” models, which include “chain-of-thought” reasoning capabilities.
This development is not entirely unforeseen. In March, OpenAI CEO Sam Altman announced on X that the company aimed to streamline its offerings, with GPT-4.5 being the last “non-chain-of-thought model.” Following that, the focus would shift to integrating the o-series and GPT models to create a versatile product tailored to users’ needs. The adoption of GPT-4o as the default model aligns with this streamlining objective.
Currently, OpenAI intends to launch o3 and o4-mini as standalone models and has postponed the release of GPT-5, which Altman anticipates will be ready “in a few months.” However, this timeline may be adjusted, as the release has experienced several delays already.
Despite its initial significance as a “pivotal moment in ChatGPT’s evolution,” GPT-4 is now considered to have fulfilled its purpose and is set to be retired.
