OpenAI has unveiled GPT-5.4, a new AI model the firm describes as its most proficient system yet for professional applications. The model integrates advanced reasoning, coding, and the capacity to autonomously run computers and software, ramping up competition for enterprise clients—an area where rival Anthropic has held a strong position.
This new model merges some capabilities OpenAI previously offered across distinct models, combining the coding prowess of GPT-5.3-Codex—the company’s top programming model—enhanced reasoning abilities, and the agentic skill to autonomously navigate desktops, browsers, and software applications.
The model will be available starting Thursday to ChatGPT Plus, Team, and Pro subscribers, as well as via OpenAI’s API. A more powerful variant, GPT-5.4 Pro, is also accessible for users needing peak performance on intricate tasks.
Alongside the model’s release, OpenAI is rolling out ChatGPT for Excel and Sheets in beta: a version of ChatGPT embedded directly into spreadsheets to create, analyze, and update complex financial models. The company is also launching new ChatGPT app integrations—including FactSet, MSCI, Third Bridge, and Moody’s—built to let teams gather market, company, and internal data into one unified workflow.
This move also places OpenAI in more direct competition with rival Anthropic, which introduced comparable products as part of its Claude for Financial Services launch in July 2025 (with expansions later that year). Both firms are racing to secure the enterprise market with tools that deliver practical value for sectors poised to adopt AI.
The latest announcement could also ignite a new wave of investor concern over AI’s effect on traditional financial data providers, many of whose stocks have already been impacted by broader worries about AI-driven disruption to enterprise software. Earlier this year, Anthropic’s Cowork plugin release prompted a widespread selloff in SaaS stocks, as markets grew nervous about the possibility of AI tools rendering legacy software providers obsolete.
New Agentic Features
One of the most notable aspects of GPT-5.4 for enterprise clients is its out-of-the-box agentic capabilities. The model can autonomously run computers and software, search for and utilize external tools as needed, and manage complex, multi-step tasks—all without developers having to build that infrastructure themselves.
OpenAI states GPT-5.4 is the company’s most factual and reliable model to date. In a blog post, the firm noted it has cut down hallucination rates: individual claims are 33% less likely to be false, and full responses are 18% less likely to contain errors compared to GPT-5.2.
“Developers don’t just need a model that writes code—they need one that approaches problems the way they do. We’re observing GPT-5.4 perform exceptionally well at logical reasoning and carrying out complex, multi-step, tool-reliant workflows,” Mario Rodriguez, Chief Product Officer at GitHub, stated about the model.
These new features position OpenAI as a more direct competitor in the increasingly crowded agentic product space—including offerings like Perplexity Computer, Microsoft’s Copilot Tasks, and OpenClaw. As seen with OpenClaw’s recent popularity, users are increasingly seeking AI systems that can manage extended workflows with minimal human intervention.
OpenAI also notes GPT-5.4 is significantly more token-efficient than its earlier models—meaning it uses fewer tokens to resolve problems. Even though it’s priced slightly higher per token than GPT-5.2, OpenAI says those efficiency improvements mean it requires fewer tokens to complete many tasks, which could balance out the higher cost for some users.