Introducing GPT-5.4 in Microsoft Foundry

We’re excited to share that OpenAI’s GPT-5.4 is now available on Microsoft Foundry! This new model is crafted to assist organisations in transitioning from planning tasks to effectively executing them in real-world environments. As AI technologies tackle more intricate workflows, maintaining consistency and follow-through is just as crucial as having advanced intelligence. GPT-5.4 enhances reasoning abilities while incorporating computer usage features to aid in automating tasks and ensuring reliable performance across various tools and multi-step processes.

This model is tailored for businesses employing AI in actual production settings, where following instructions, maintaining consistency, and keeping context intact are vital for success. It brings together improvements in reasoning, coding, and task management, allowing AI systems to not only strategize but also carry out tasks with fewer disruptions and less need for hands-on control.

Compared to its predecessors, GPT-5.4 focuses on stability throughout extended interactions, allowing teams to adopt agentic AI solutions with more assurance in daily operations.

Here are some noteworthy advancements in GPT-5.4 aimed at achieving production-grade AI:

• Consistent reasoning over time: Helps sustain intent during prolonged interactions.
• Better instruction alignment: Minimises the need for frequent adjustments and supervision.
• Reduced latency: Ensures swift, real-time responses for efficient workflows.
• Integrated computer use features: Supports structured management of tools, data access, code execution, and smooth agent transitions.
• Reliably invoking tools: Lessens the necessity for oversight and adjustments.
• Enhanced quality of generated content: Produces structured documents, spreadsheets, and presentations more consistently.

All these improvements lead to AI systems that behave more predictably, even as the complexity and length of tasks expand.

GPT-5.4 provides tangible benefits across various production scenarios where reliability and follow-through are essential, including:

• Agent-driven workflows: Such as customer support, research assistance, and automating business processes.
• Enterprise knowledge work: Encompassing document drafting, data analysis, and ready-to-present outputs.
• Developer tasks: Including code generation, debugging support, and user interface setup.
• Extended reasoning activities: Where logical consistency must be maintained across lengthy interactions.

With GPT-5.4, teams can expect fewer disruptions, smoother workflows, and more predictable results in their production environments.

The premium variant, GPT-5.4 Pro, is designed for scenarios where a deeper analytical approach is more crucial than speed. Its additional features comprise:

• Multi-path reasoning evaluation: Allowing exploration of various approaches before deciding on a final solution.
• Increased analytical depth: Aiding in the resolution of complex issues or trade-offs.
• Improved consistency in reasoning: Especially beneficial in sustained analytical tasks.
• Better decision support: Focusing on thoroughness and accuracy over speed.

Generally, organisations opt for GPT-5.4 Pro when in-depth analysis is needed, such as in scientific research or complex problem-solving, while GPT-5.4 is ideal for tasks prioritising reliable execution and consistent follow-through.

Both GPT-5.4 and GPT-5.4 Pro are accessible through Microsoft Foundry, which offers essential operational controls for responsible AI deployment in production. Foundry facilitates policy enforcement, monitoring, version management, and audit trails, enabling teams to oversee their AI systems effectively throughout their lifecycle.

By integrating GPT-5.4 via Microsoft Foundry, organisations can seamlessly infuse advanced AI capabilities into their existing setups while meeting security, compliance, and operational requirements from the outset.

Customer Highlight

GPT-5.4 raises the standard for production-ready AI with its superior reasoning and reliable execution. Through enterprise-level deployment on Microsoft Foundry, businesses can confidently move past experimentation and construct AI systems capable of tackling complex tasks at scale. Enhanced computer use capabilities are set to be revealed shortly after the launch.

The pricing for GPT-5.4 in Microsoft Foundry is $2.50 per million input tokens, $0.25 per million cached input tokens, and $15.00 per million output tokens, available at launch in Standard Global and Standard Data Zone (US), with more options forthcoming. For GPT-5.4 Pro, the cost is $30.00 per million input tokens and $180.00 per million output tokens, also available in Standard Global at launch.

Begin crafting agents for real-world tasks. Dive into GPT-5.4 in Microsoft Foundry today!