Release Notes

Continuing to retire older models with limited usage in ChatGPT. OpenAI o3 will be retired from ChatGPT on August 26, 2026 following a 90-day sunset period, and GPT-4.5 will be retired from ChatGPT on June 27, 2026 following a 30-day sunset period. These models are currently available to paid users only via model settings. These changes apply to ChatGPT only; there are no changes to the API.

Updated GPT-5.5 Instant in ChatGPT and the API to improve response style and quality. Responses are now easier to read, more natural in everyday conversations, and better paced in practical help tasks, with fewer overly long or bullet-heavy responses.

Canvas is no longer available in GPT-5.5 Instant or GPT-5.5 Thinking. Writing and coding functionality is now supported directly in chat responses through writing blocks and code blocks. Paid users can continue using canvas for a limited time through legacy models until those models are sunset.

Rolling out GPT-5.4 mini in ChatGPT. GPT-5.4 mini is available to Free and Go users via the "Thinking" feature in the + menu. For all other users, GPT-5.4 mini is available as a rate limit fallback for GPT-5.4 Thinking.

For Plus, Pro, and other paid users, GPT-5.4 mini will be used as a fallback for GPT-5.4 Thinking when rate limits are reached. Enterprise customers will retain the option to default Auto routing to GPT-5.4 mini, if preferred.

GPT-5.4 mini will not appear as a selectable model in the model picker, and GPT-5 Thinking mini will be retired as a selectable option in 30 days.

Rolling out an update to GPT-5.3 Instant that improves follow-up tone and reduces teaser-style phrasing in responses (e.g., "If you want…", "You'll never believe…", "I can tell you these three things that…").

As of March 11, 2026, GPT-5.1 models are no longer available in ChatGPT. This applies to GPT-5.1 Instant, GPT-5.1 Thinking, and GPT-5.1 Pro. Existing conversations that used GPT-5.1 will automatically continue on the corresponding current model: GPT-5.3 Instant, GPT-5.4 Thinking, or GPT-5.4 Pro.

GPT-5.4 brings together the best of recent advances in reasoning, coding, and agentic workflows into a single frontier model. It incorporates industry-leading coding capabilities while improving how the model works across tools, software environments, and professional tasks involving spreadsheets, presentations, and documents.

In ChatGPT, GPT-5.4 Thinking can now provide an upfront plan of its thinking, so you can adjust course mid-response while it's working and arrive at a final output that's more closely aligned with what you need without additional turns. GPT-5.4 Thinking improves deep web research, particularly for highly specific queries, while better maintaining context for questions that require longer thinking. It also has improved context window management that supports its ability to think for longer.

Retired GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini from ChatGPT. Also retiring GPT-5 (Instant and Thinking), as previously announced. There are no API changes at this time.

Making an update to GPT-5.2 Instant in ChatGPT and the API that improves response style and quality. Responses are now more measured and grounded in tone, in a way that's more contextually appropriate to the conversation. The model also tends to output clearer, more relevant answers to advice-seeking and how-to questions, more reliably placing the most important info upfront.

Launched GPT-5.3-Codex, the most capable agentic coding model yet. This model is the first that combines Codex + GPT-5 training stacks, bringing together best-in-class code generation, reasoning, and general-purpose intelligence in one unified model. It's approximately 25% faster, sets new highs on key benchmarks, and marks a step-change from code generation to a general-purpose coding agent you can actively steer while it works.

Series of updates to thinking time settings for GPT-5.2 Thinking:

• January 10, 2026: Lowered Standard and Light thinking time as users prefer faster responses. Extended thinking setting was unintentionally changed to be lower, which was then fixed.
• February 3, 2026: Made another small reduction to Standard thinking time based on testing.
• February 4, 2026: Restored Extended thinking level for GPT-5.2 Thinking to its prior setting, correcting the inadvertent reduction from January.

Thinking time is not directly comparable across different models. Each model is tuned independently to what works best for users.

Updating GPT-5.2 Instant's default personality to be more conversational and better at adapting its tone contextually, making exchanges feel smoother and more natural. Users can still select a different base style and tone for ChatGPT, along with tuning characteristics like warmth and emoji use, within the Personalization menu in settings.

Updated the Model Spec to strengthen and more clearly codify the principles that reflect how OpenAI builds experiences for teen users.

New section: Under-18 (U18) Principles - ChatGPT's new Under-18 Principles builds on existing safety rules, adding age-appropriate guidance for teens aged 13-17. This clarifies how rules are intended to apply in teen conversations, recognizing that teens benefit from clearer boundaries, reduced exposure to potentially harmful content, and stronger real-world support when risks arise. The assistant should refuse to participate in self-harm, sexualized or violent immersive roleplay, dangerous activities, substance misuse, or efforts to conceal harm. When credible risks arise, the model should prioritize prevention and early interventions, offer safer alternatives, and encourage involvement of parents, guardians, and other trusted adults or professionals.

Also includes minor edits and clarifications for consistency and readability throughout the document.

Introducing a new GPT-5-Codex-Mini model option to Codex CLI and the IDE Extension. The model is a smaller and more cost-effective version of GPT-5-Codex that provides up to 4x more usage as part of your ChatGPT subscription.

Starting today, Codex in both the CLI and IDE Extension will automatically offer to switch to GPT-5-Codex-Mini when you reach 90% of your 5-hour usage limit to help you work longer without interruptions.

GPT-5.1-Codex-Max is a new frontier agentic coding model built for long-running, project-scale work. It's faster, more capable, and more token-efficient than GPT-5.1-Codex, using compaction to work coherently across multiple context windows. Available in Codex surfaces including the CLI, IDE extension, cloud, and code review. Rates are the same as for GPT-5.1-Codex.

Updating GPT-5 Instant to better recognize and support people in moments of distress. The model is trained to more accurately detect and respond to potential signs of mental and emotional distress, guided by mental health experts. It helps ChatGPT de-escalate conversations and point people to real-world crisis resources when appropriate, while using language that feels supportive and grounding.

GPT-5 Instant now performs just as well as GPT-5 Thinking on distress-related questions. When GPT-5 Auto or a non-reasoning model is selected, sensitive conversations will be routed to GPT-5 Instant to more quickly provide helpful responses.

GPT-5-codex is now available in the Responses API, in addition to codex surfaces. Note: GPT-5-Codex is not currently supported in ChatGPT.

Adding GPT-5-codex, a GPT-5 variant optimized for agentic coding in Codex. It's available everywhere you use Codex: default for cloud tasks and code review, and selectable for local workflows via the Codex CLI and IDE extension. Use GPT-5-codex for coding-focused work in Codex, or Codex-like environments; use GPT-5 for general, non-coding tasks.

In day-to-day use, GPT-5-codex supports fast interactive edits and can run independently on longer tasks when needed. For frontend/UI work, it accepts images or screenshots alongside text as input. Note: GPT-5-Codex is not currently supported in ChatGPT.

Updated the Model Spec to better reflect how OpenAI's systems are evolving, focusing on strengthening clarity and guardrails as models move beyond chat into more agentic use cases.

Updated authority levels - The top authority level has been renamed from Platform to Root and elevated above System, making clear which parts of the Model Spec cannot be overridden in any conversation. New authority order: Root → System → Developer → User → Guideline.

Agentic principles - Added principles for agents that can take actions in the world: (1) Act within an agreed-upon scope of autonomy, and (2) Control and communicate side effects.

Other notable changes - Improvements to the Chain of Command with a new No other objectives section; expanded context on OpenAI's goals for safe model behavior in the Overview; expanded principles and examples for default model personality; clarified language around system and developer message confidentiality; and updated refusal style to safe completion for more helpful and transparent model responses around safety boundaries.

GPT-5 is slowly rolling out to all users on ChatGPT Plus, Pro, Team, and Free plans worldwide across web, mobile, and desktop. GPT-5 will be available to ChatGPT Enterprise and Edu plans soon.

GPT-5 in ChatGPT is the next flagship model and the new default for all logged-in users. It simplifies ChatGPT to a single auto-switching system that brings together the best of previous models into a smart, fast model.

GPT-5 is available to all ChatGPT Tiers. Users on Paid tiers (Plus, Pro, and Team) have access to the model picker, which enables you to manually select GPT-5 or GPT-5 Thinking. Pro and Team tier users have access to GPT-5 Thinking Pro, which takes a bit longer to think but delivers the accuracy needed for complex tasks.

Releasing two open-weight reasoning models, gpt-oss-120b and gpt-oss-20b. Designed for teams that want to run and customize models on their own infrastructure or with hosting providers, these text-only models support common developer patterns like function calling and structured outputs.