OpenAI described its new GPT as “the most capable and efficient frontier model for professional work” and claimed that it brings together the company’s advances in reasoning, programming and agentic workflows into one unified model, Gizmodo reports.
GPT-5.4 is available from Thursday and will be usable in ChatGPT, Codex and through the OpenAI API. GPT-5.4 Thinking will be available to Plus, Teams and Pro users, and GPT-5.4 Pro will be accessible via API as well as to ChatGPT Enterprise and Edu subscribers.
According to OpenAI, GPT-5.4 is the first general-purpose model released by the company that has native computing capabilities, meaning it can operate autonomously in various applications on a computer on behalf of the user.
OpenAI stated that the model can write code to operate and execute tasks on computers, as well as issue keyboard and mouse commands to navigate the operating system.
The release of GPT-5.4, at a delicate moment for OpenAI
In terms of more general uses, such as asking questions, OpenAI says the newest model delivers better results in this department as well. The company claims that individual responses generated by GPT-5.4 are 33% less likely to contain errors compared to GPT-5.2 responses, and the new model is 18% less likely to make mistakes overall.
The company also said that the occurrence of “hallucinations” is less likely in the case of GPT-5.4, although in this chapter it did not give a percentage.
Its launch comes after ChatGPT lost around 1.5 million users amid a boycott movement in the US.
This comes as OpenAI announced it would work with the Pentagon after rival company Anthropic demanded assurances from the Pentagon that the models it provided would not be used for surveillance of Americans or fully autonomous weapons systems.
The Pentagon denied the request and the Trump administration banned the use of Anthropic models, the company behind the Claude chatbot, in any branch of the federal government.
PHOTO: Rokas Tenys / Dreamstime.
