OpenAI Unveils GPT-5.4, a Professional‑Focused AI Model
Key Points
- OpenAI released GPT-5.4, a model optimized for professional tasks like coding, data analysis, and presentation design.
- The model introduces native computer‑use capabilities, improving mouse and keyboard interactions across multiple apps.
- In ChatGPT, GPT-5.4 powers the new Thinking mode, outlining response plans before generation.
- OpenAI says GPT-5.4 is 18% less likely to produce errors than GPT-5.2, making it its most factual model yet.
- The model is targeted at enterprise customers and developers, not available to free or Plus users.
- API token pricing increased to $2.50 per million input tokens, reflecting higher efficiency but higher cost.
- The release underscores OpenAI’s strategic shift toward productivity‑focused revenue streams.
OpenAI announced GPT-5.4, its latest frontier model built for professional tasks such as coding, data analysis, and presentation creation. The model adds native computer‑use abilities, allowing smoother mouse and keyboard interaction across multiple applications. In ChatGPT, GPT-5.4 becomes the default for the Thinking mode, outlining its plan before generating responses and supporting more precise web research. OpenAI positions the model as its most factual to date, citing an 18% reduction in error likelihood versus GPT-5.2. While priced higher for API tokens and limited to enterprise and developer customers, the release signals OpenAI’s shift toward productivity‑oriented revenue streams.
Introducing GPT-5.4
OpenAI released a new large‑language model called GPT-5.4, describing it as the company’s most capable frontier model for professional work. The model is engineered for tasks such as coding, data analysis, and the creation of presentations with stronger, more varied aesthetics. OpenAI also notes that GPT-5.4 makes more effective use of its image‑generation tools, enhancing visual outputs.
Native Computer‑Use Capabilities
GPT-5.4 is the first OpenAI model built with native computer‑use capabilities. This allows the system to issue mouse and keyboard commands more accurately and to navigate desktop environments better than its predecessor, GPT-5.2. The model can operate across several applications simultaneously, a step forward for multi‑app workflows.
ChatGPT Integration and User Interaction
Within ChatGPT, GPT-5.4 will be the default model for the new "Thinking mode." In this mode, the system first outlines how it intends to address a request, giving users the chance to adjust the plan before the response is generated. OpenAI also highlights improved web‑research abilities, especially for "highly specific" queries, promising higher‑quality answers that arrive faster and stay relevant to the task at hand.
Accuracy and Factuality
OpenAI claims GPT-5.4 is its most factual model yet, stating it is 18 percent less likely to produce erroneous responses compared with GPT-5.2. The company frames this improvement as part of a broader effort to deliver more reliable AI assistance.
Availability, Pricing, and Target Audience
GPT-5.4 will be accessible in ChatGPT when users select Thinking mode and can also be chosen as "GPT-5.4 Pro" from the model picker. The model is not offered to Free, Go, or Plus subscribers; instead, it is aimed at enterprise customers and developers who rely on OpenAI’s Codex application. For API users, GPT-5.4 is described as the most token‑efficient reasoning model to date, though token costs have risen to $2.50 per million input tokens, up from $1.75 for GPT-5.2.
Strategic Shift Toward Professional Productivity
The launch reflects OpenAI’s growing focus on professional productivity tools. The shift follows Microsoft’s decision last September to add Anthropic’s models to Copilot 365 after finding that Claude performed well on spreadsheet and presentation tasks. Reporting suggests OpenAI generates about $25 billion in annualized revenue but remains far from profitability, with $1.4 trillion in data‑center commitments and continued reliance on investor funding. By emphasizing productivity‑oriented features, OpenAI appears to be seeking a sustainable business model within the enterprise space.