OpenAI announced GPT 5.3 in the chat just a couple of days ago. GPT 5.4 is the thinking model. It has advanced reasoning, coding, and agentic workflows. It is a faster, more efficient model, so you can use it with fewer tokens. It has better deep web research and context retention. You will be able to interrupt and redirect the model in the middle of its work.
Interrupting and Adding Details in GPT-5.4 Thinking
GPT 5.4 can provide an upfront plan of its thinking, allowing you to make adjustments before it is done. It has a 1M token context. It uses fewer tokens to solve problems in comparison to GPT 5.2.
[HT]
