Robots Atlas>ROBOTS ATLAS
GPT-5.4 Pro
AI Modelsโ€บGPT

GPT-5.4 Pro

5.4 Proย ยทย Family: GPT
The highest-quality GPT-5.4 variant, using increased compute to handle highly complex problems; available in ChatGPT and via API.
โœ“ Activeโ˜… FeaturedLLMMultimodalReasoning modelVisionTool-using model๐Ÿ“ GPT
Context window
1.05M tokens
tokens
Parameters
Nieujawnione
parameters
Release date
5 March 2026
๐ŸขOpenAIProducer

Overview

GPT-5.4 Pro is a variant of GPT-5.4 designed for the most demanding tasks, using increased compute to deliver more precise and consistently higher-quality responses. The model supports text and image input and generates text output. In the API, it is available primarily through the Responses API and supports reasoning.effort values of medium, high, and xhigh. In ChatGPT, the Pro variant is available to Pro, Business, Enterprise, and Edu plan subscribers.

Classification
LLMMultimodalReasoning modelVisionTool-using model
Family: GPT
Applications
CodingDocument analysisQ&A / Question answeringResearch assistanceSearch assistanceSpreadsheet analysisPresentation generationSummarization
Key parameters
๐Ÿ“ Context: 1.05M tokens
๐Ÿงฉ Parameters: Nieujawnione
โœ“ Tools
๐Ÿ“ฅ Input: text, image

Technical specification

Context window
1.05M tokens
tokens
Parameters
Nieujawnione
parameters
License
Proprietary (OpenAI)
Hardware requirements
No local hardware requirements for the end user. The model is provided as a service via the OpenAI API and in ChatGPT. In practice, only a client environment and access to OpenAI services are required. Local self-hosting is not supported.
Features:โœ“ Tool use
Modalities
โฌ‡ Input
textimage
โฌ† Output
text

Capabilities and applications

Native model capabilities
Reasoning
The model's ability to reason logically and solve complex problems.
Category: reasoning
Multi-step reasoning
Carrying out multi-step chains of reasoning across long, complex tasks.
Category: reasoning
Long context
Maintaining coherence and focus across very long input context.
Category: language
Coding
Generating, analysing and modifying source code.
Category: coding
Function Calling
Category: planning
Structured output
Producing data in structured formats such as JSON.
Category: structured_generation
Audio understanding
Category: audio
Image understanding
Analysing and interpreting the content of images.
Category: vision
Video Understanding
Category: video
Chart understanding
Reading and interpreting charts, tables and diagrams.
Category: vision
Diagram reasoning
Category: reasoning
OCR
Recognising text within images and documents.
Category: vision
Multilingual
Understanding and generating text in many languages.
Category: language
Planning
Forming and executing action plans for complex tasks.
Category: planning
Streaming output
Category: reasoning
Interleaved Multimodal Input
Category: reasoning
Multimodal understanding
Category: multimodal
Application domains
CodingDocument analysisQ&A / Question answeringResearch assistanceSearch assistanceSpreadsheet analysisPresentation generationSummarizationWriting assistance

Benchmark results

1 benchmark
BrowseComp
accuracy ยท agentic web browsing
89.3%
๐Ÿ“… 5 Mar 2026๐Ÿ“„ OpenAI โ€“ Introducing GPT-5.4
GPT-5.4 Pro sets a new SOTA on BrowseComp, according to OpenAI.

Pricing

Technical architecture

Deployment and security

๐Ÿ”’ Security / Enterprise
โœ“ Verified enterprise information

Public security information for GPT-5.4 Pro derives from OpenAI's security policies and architecture for the API platform and ChatGPT Enterprise/Business/Edu products. The vendor describes security primarily at the service level, rather than in a separate model-specific document for the Pro variant.

For GPT-5.4 Pro, security should be modeled as platform security inherited from the OpenAI environment. This is the most accurate and substantively correct framing.
Updated: 15 Mar 2026โ†— Security documentation

Sources and related pages