The model is rolling out to Plus, Pro, Business, and Enterprise users across ChatGPT and Codex, with API access expected soon. GPT-5.5 Pro is also being made available to Pro, Business, and Enterprise users.
What’s new in GPT-5.5
The biggest change in GPT-5.5 is how it handles instructions. Instead of requiring step-by-step guidance, users can provide a loosely structured request, and the model will plan and complete the workflow.
This marks a shift from prompt-driven interactions to more autonomous behaviour, where the model takes on a larger share of the task.
OpenAI said GPT-5.5 performs well in areas such as coding, research, data analysis, and document creation, where tasks involve multiple stages. The model is also capable of moving across tools and software environments to complete workflows instead of responding in isolated steps.
The company added that GPT-5.5 maintains response speeds similar to GPT-5.4 while offering improved capability. It also uses fewer tokens for similar tasks, which may reduce the need for repeated prompts.
Different modes in ChatGPT
GPT-5.5 is available in different modes designed for varying levels of complexity.
-
GPT-5.5: Focuses on understanding intent, gathering information, and converting raw inputs into structured outputs. -
GPT-5.5 Thinking: Designed for complex problems, offering structured reasoning and concise responses for tasks such as coding, research, and analysis. -
GPT-5.5 Pro: Available to higher-tier users, aimed at more demanding workloads with improved accuracy and depth, particularly in business, legal, education, and data-related tasks.
These modes are configurations of the same model rather than separate systems.
What changes in Codex
In Codex, GPT-5.5 is geared towards productivity and content generation.
OpenAI said the model performs better at:
-
Generating documents, spreadsheets, and presentations -
Handling operational research and spreadsheet modelling -
Converting unstructured inputs into structured plans
The update builds on recent Codex features, including improved document and slide generation, richer previews, in-app browser use, image generation, and auto review tools.
Safeguards and risk controls
OpenAI said GPT-5.5 includes updated safeguards, particularly for sensitive areas such as cybersecurity.
The company has introduced stricter controls for high-risk queries and expanded monitoring systems to detect misuse.
Its capabilities in areas such as cybersecurity and biology are classified as high under OpenAI’s preparedness framework, indicating increased capability compared to earlier models, though not the highest risk level.
Availability
GPT-5.5 is rolling out across ChatGPT and Codex for Plus, Pro, Business, and Enterprise users. GPT-5.5 Pro is limited to Pro, Business, and Enterprise tiers.
API access is expected soon, with pricing likely to be higher than GPT-5.4. However, OpenAI said improved efficiency may offset costs by reducing token usage per task.
