OpenAI has introduced GPT-5.4, the latest version of its flagship AI model, bringing upgrades in reasoning, coding, and task automation. The model is rolling out across ChatGPT, the API, and developer tools, with different variants tailored for everyday users and enterprise applications.
A key addition in GPT-5.4 is its ability to interact more directly with computers. The model can analyze screenshots, navigate web browsers, and execute keyboard and mouse commands to complete tasks across applications and services.
This enables GPT-5.4 to manage multi-step workflows that previously required human involvement. The development represents a move toward more autonomous AI agents capable of carrying out real-world tasks.
OpenAI stated that GPT-5.4 improves its handling of complex research queries. The model can conduct multiple rounds of information gathering and combine results into structured responses.
According to the company, GPT-5.4 is its most factual model to date, reducing false claims by approximately 33 percent compared to GPT-5.2.
OpenAI also introduced GPT-5.4 Thinking within ChatGPT. This mode is designed for more challenging prompts and displays an outline of the model’s reasoning process as it works through a problem.
Users can modify instructions during the response, allowing them to guide the AI without restarting the session.
GPT-5.4 supports longer and more complex tasks, retaining context across extended workflows. These enhancements may benefit tools such as OpenAI Codex, where the model can automate large development tasks.
The rollout has begun for ChatGPT users on the web and Android, with iOS support expected soon. OpenAI is also offering GPT-5.4 Pro for enterprise and academic users who require higher performance for complex workloads.
Get the latest tech news, telecom insights, and product launches wherever you prefer.
Add ProPakistani to Preferred Sources and see more of our stories in Google Search and Top Stories.