OpenAI announces that it will release 'o3' and 'o4-mini' within weeks, but postpones the release of 'GPT-5'

OpenAI CEO Sam Altman spoke about upcoming product releases on his X account, revealing the approximate release dates for the unannounced models 'o3', 'o4-mini', and 'GPT-5'.

change of plans: we are going to release o3 and o4-mini after all, probably in a couple of weeks, and then do GPT-5 in a few months.

there are a bunch of reasons for this, but the most exciting one is that we are going to be able to make GPT-5 much better than we originally…

— Sam Altman (@sama) April 4, 2025

CEO Altman often mentions product release schedules on his X account. On February 13, 2025, he announced that GPT-4.5 will be released within a few weeks and GPT-5 will be released within a few months. In addition, he also announced that o3 , the full version of o3-mini , will be released in an integrated form with GPT-5.

OpenAI CEO mentions the planned release of 'GPT-4.5' and 'GPT-5', GPT-4.5 is the last non-inference model and GPT-5 integrates multiple technologies - GIGAZINE

Subsequently, on February 27, 2025, GPT-4.5 was released as CEO Altman had predicted.

OpenAI releases GPT-4.5, calling it the 'biggest and best conversation model ever' - GIGAZINE

Newly, on April 4, 2025, CEO Altman said on his X account, 'Plan change: After all, we have decided to release o3 and o4-mini. o3 and o4-mini will be released in a few weeks. And GPT-5 will be released in a few months.' In other words, o3 will be released as a standalone model rather than integrated with GPT-5, and the next-generation o4-mini will also be released prior to the release of GPT-5.

change of plans: we are going to release o3 and o4-mini after all, probably in a couple of weeks, and then do GPT-5 in a few months.

there are a bunch of reasons for this, but the most exciting one is that we are going to be able to make GPT-5 much better than we originally…

— Sam Altman (@sama) April 4, 2025

Regarding the reason for delaying the release of GPT-5, CEO Altman said, 'There are many reasons, but the most exciting one is that we can make GPT-5 much better than we originally thought. We also realized that integrating everything smoothly was more difficult than we thought. In addition, we expect unprecedented demand for GPT-5, so we want to ensure we have enough capacity.'

The lack of computing resources is a serious problem for OpenAI, and the release of GPT-4.5 was delayed due to a lack of GPUs. In March 2025, the company released a high-performance image generation function called ' 4o Image Generation ', but restrictions were placed on its use due to a lack of GPUs.

OpenAI says ChatGPT's new image generation function is so popular that it 'melts the GPU', forcing temporary restrictions - GIGAZINE