Skip to main content

Safety

Safety at every step

We believe in AI’s potential to make life better for everyone, which means making it safe for everyone

Teach

Test

Share

Safety doesn’t stop

Building safe AI isn’t one and done. Every day is a chance to make things better. And every step helps anticipate, evaluate, and prevent risk.

How we think about safety and alignment

Leading the way in safety

We collaborate with industry leaders and policymakers on the issues that matter most.

Child safety

Private information

Deep fakes

Bias

Elections

Conversations with OpenAI researchers

Get inside OpenAI with our series that breaks down a range of topics around safety and more.

Latest news on safety

Parental-Controls_Art Card 1080x1080

Introducing parental controls

ProductSep 29, 2025

PFv2 artcard 1.1

Our updated Preparedness Framework

PublicationApr 15, 2025

Update on disrupting > Cover

An update on disrupting deceptive uses of AI

SafetyOct 9, 2024

safety-blog-cover-02

OpenAI safety practices

SafetyMay 21, 2024

Go deeper on safety

Explore the safety evaluations hub

Sora 2 System Card

Sora 2’s advanced capabilities require consideration of new potential risks, including nonconsensual use of likeness or misleading generations. To address these, we worked with internal red teamers to identify new challenges and inform corresponding mitigations. We’re taking an iterative approach to safety, focusing on areas where context is especially important or where risks are still emerging and are not fully understood.

Addendum to GPT-5 system card: GPT-5-Codex

This addendum outlines the comprehensive safety measures implemented for GPT‑5-Codex. It details both model-level mitigations, such as specialized safety training for harmful tasks and prompt injections, and product-level mitigations like agent sandboxing and configurable network access.

GPT-5 System Card

In this system card, we label the fast, high-throughput models as gpt-5-main and gpt-5-main-mini, and the thinking models as gpt-5-thinking and gpt-5-thinking-mini. In the API, we provide direct access to the thinking model, its mini version, and an even smaller and faster nano version of the thinking model, made for developers (gpt-5-thinking-nano).

Addendum to OpenAI o3 and o4-mini system card: OpenAI o3 Operator

This addendum to o3 and o4-mini system card describes the release of Operator, a Computer Using Agent capable of interacting with the web through its own browser. Operator is now powered by a o3-based model (instead of GPT-4o) with enhanced safety fine-tuning for computer use.

Addendum to OpenAI o3 and o4-mini system card: Codex

This addendum to o3 and o4-mini system card describes Codex, a cloud-based coding agent powered by codex-1, a fine-tuned version of OpenAI’s o3 model tailored specifically for software engineering tasks.

OpenAI o3 and o4-mini System Card

This is the first launch and system card to be released under Version 2 of our Preparedness Framework. OpenAI’s Safety Advisory Group (SAG) reviewed the results of our Preparedness evaluations and determined that OpenAI o3 and o4-mini do not reach the High threshold in any of our three Tracked Categories: Biological and Chemical Capability, Cybersecurity, and AI Self-improvement.

Addendum to GPT-4o System Card: 4o image generation

This addendum to the GPT‑4o system card describes the marginal risks we’ve focused on, and the work we have done to address them.

GPT-4.5 System Card

This system card outlines how we built and trained GPT‑4.5, evaluated its capabilities, and strengthened safety, following OpenAI’s safety process and Preparedness Framework.

Deep research System Card

This report outlines the safety work carried out prior to releasing deep research including external red teaming, frontier risk evaluations according to our Preparedness Framework, and an overview of the mitigations we built in to address key risk areas.

OpenAI o3-mini System Card

This report outlines the safety work carried out for the OpenAI o3-mini model, including safety evaluations, external red teaming, and Preparedness Framework evaluations.

Operator System Card

This report outlines the safety work carried out prior to releasing Operator including external red teaming, frontier risk evaluations according to our Preparedness Framework, and an overview of the mitigations we built in to address key risk areas.

Sora System Card

This system card outlines the resulting mitigation stack, external red teaming efforts, evaluations, and ongoing research to refine these safeguards further.

OpenAI o1 System Card

This report outlines the safety work carried out prior to releasing OpenAI o1 and o1-mini, including external red teaming and frontier risk evaluations according to our Preparedness Framework.

GPT-4o System Card

This system card takes a detailed look at speech-to-speech while also evaluating text and image capabilities.