GPT-5.4 Is Now Available in Microsoft Foundry — Here's What It Means for Partners

OpenAI's GPT-5.4 is generally available in Microsoft Foundry. This isn't just another model upgrade — it's built for production workloads where consistency and follow-through matter as much as raw intelligence.

What's New

GPT-5.4 is designed for organizations running AI in real production environments. The key advances:

- More consistent reasoning across multi-turn, multi-step interactions — the model maintains intent over longer workflows
- Better instruction alignment — less prompt tuning, less manual oversight
- Improved latency for responsive, real-time use cases
- Built-in computer use capabilities — structured orchestration of tools, file access, data extraction, code execution, and agent handoffs (coming shortly after launch)
- More reliable tool invocation — fewer mid-workflow failures
- Higher-quality generated artifacts — documents, spreadsheets, and presentations with more consistent structure

In short: GPT-5.4 is about AI that finishes what it starts.

Two Variants

GPT-5.4 — the production workhorse. Best for agent-driven workflows, enterprise knowledge work, developer tooling, and any scenario where reliable execution matters.

GPT-5.4 Pro — a premium variant for deeper analytical work. It evaluates multiple reasoning paths before selecting a response, making it suited for scientific research, complex trade-off analysis, and scenarios where thoroughness outweighs speed.

Pricing

Model	Input	Cached Input	Output
GPT-5.4	$2.50/M tokens	$0.25/M tokens	$15.00/M tokens
GPT-5.4 Pro	$30.00/M tokens	—	$180.00/M tokens

GPT-5.4 launches in Standard Global and Standard Data Zone (US), with more deployment options coming. GPT-5.4 Pro is available in Standard Global.

Why This Matters for Partners

If your customers are building or evaluating AI agents — for customer support, business process automation, document generation, or code workflows — GPT-5.4 in Microsoft Foundry gives them a production-grade option with enterprise controls built in. Policy enforcement, monitoring, version management, and auditability are all handled through Foundry.

This is the model you point to when a customer says: "We've been experimenting with AI. How do we actually deploy it?"

Get started: Microsoft Foundry Model Catalog