Termsby GlyphOpenAI data terms: training, retention, opt-out
Model APIs, ChatGPT · 5 pages watched · last change Nov 2, 2026
What the terms say
Does it train on API and business data?
No, unless you opt in(Protective)
“By default, we don't use inputs or outputs from ChatGPT Business, ChatGPT Enterprise, ChatGPT Edu, or our API to improve our models.”
How your data is used to improve model performance On the page Oct 26, 2026
Does it train on consumer app data?
Yes, unless you opt out(Uses your data)
“When you use our services for individuals, such as ChatGPT and Codex, we may use your content to train our models.”
How your data is used to improve model performance No longer on the page; being re-read
How long are API inputs kept?
Abuse logs up to 30 days(Depends)
“By default, abuse monitoring logs are generated for all API feature usage and retained for up to 30 days, unless longer retention is required by law, or is reasonably necessary to protect our services or any third party from harm.”
Data controls in the OpenAI platform On the page Nov 2, 2026
How do you opt out of training?
Settings or privacy portal(Protective)
“To opt out, turn off Improve the model for everyone under Settings > Data controls in ChatGPT, or select Do not train on my content in our Privacy Portal .”
How your data is used to improve model performance On the page Oct 26, 2026
Is zero data retention available?
On approval(Depends)
“Currently, these controls are subject to prior approval by OpenAI and acceptance of additional requirements.”
Data controls in the OpenAI platform On the page Nov 2, 2026
Is there a public subprocessor list?
Published list(Protective)
“OpenAI engages the following entities to provide processing activities for Customer Data (as defined in the OpenAI Data Processing Agreement).”
Sub-processor list On the page Oct 26, 2026
Where is data processed and stored?
Regions for eligible customers(Depends)
“Data residency controls are a project configuration option that allow you to configure the location of infrastructure OpenAI uses to provide services.”
Data controls in the OpenAI platform On the page Nov 2, 2026
Changes
Data controls in the OpenAI platform
Added 10 paragraphs and removed 8.New text: “By default, abuse monitoring logs are retained for up to 30 days.”
View the diff
Data controls in the OpenAI platformRemoved 5 paragraphs and added 4.New text: “Image generation is Zero Data Retention compatible when using gpt-image-2.5-sunburst , gpt-image-2.5-sunburst-2026-09-08 , gpt-image-2.5-flare , gpt-image-2.5-flare-2026-09-08 , gpt-image-2 , gpt-image-1.5 , and gpt-image-1-mini .”
View the diff
Sub-processor listAdded 4 paragraphs.New text: “Processing is performed at the data center (opens in a new window) that is closest to the End User.”
View the diff
Privacy PolicyAdded 25 paragraphs and removed 9.New text: “Usage Data : We collect information about your use and activity across the Services, such as the types of content that you view or engage with, the features you use and the actions you take, when you submit feedback to a model response, the”
View the diff
How your data is used to improve model performanceAdded 3 paragraphs and removed 2; the clause quoted for trains on app data is no longer on the page.New text: “When you use our services for individuals (Free, Plus, Pro) , we may use your content to train our models. You can choose whether your conversations help improve our models. To opt out, turn off Improve the model for everyone under Settings”
View the diff
Data controls in the OpenAI platformAdded 9 paragraphs and removed 3.New text: “By default, abuse monitoring logs are retained for up to 30 days. Eligible customers can use Zero Data Retention, subject to the limitations below.”
View the diff
Privacy PolicyAdded 25 paragraphs and removed 9.New text: “Usage Data : We collect information about your use and activity across the Services, such as the types of content that you view or engage with, the features you use and the actions you take, when you submit feedback to a model response, the”
View the diff
Data controls in the OpenAI platformRemoved 8 paragraphs and added 7.New text: “Fast mode is not available with EU data residency for GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol, or GPT-6 Luna. GPT-6.1 Sol, GPT-6 Sol, and GPT-6 Luna support EU data residency with Standard, Flex, and Batch processing. Ultrafast mode supports US”
View the diff
Privacy PolicyRemoved 30 paragraphs and added 14.New text: “For individuals in the Republic of Korea, please refer to the Korea Addendum .”
View the diff
How your data is used to improve model performanceAdded 18 paragraphs and removed 12; the clause quoted for trains on app data is no longer on the page.OpenAI rewrote this help article between Sep 21 and Sep 23, 2026. The three sentences Terms quotes were reworded, not dropped, and the answers still hold: business and API data is not used to improve models by default (the list now also names ChatGPT Edu), app conversations may be used to train unless you opt out, and the opt-out settings are the same. New: a temporary chat you save becomes a regular chat and follows your model-improvement setting. Terms now quotes the new wording.Answers re-checked: Trains on API data, Opt-out
View the diff
Data controls in the OpenAI platformAdded 59 paragraphs and removed 26.New text: “Abuse monitoring logs: Logs generated from your use of the platform, necessary for OpenAI to enforce our Usage Policies and agreements and mitigate harmful uses of AI.”
View the diff
Privacy PolicyRemoved 11 paragraphs and added 10.New text: “At OpenAI, our mission is to ensure that artificial general intelligence benefits everyone. We build tools like ChatGPT to help people learn, create, and solve problems. We at OpenAI (together with our affiliates, "OpenAI", "we", "our" or "”
View the diff
How your data is used to improve model performanceRewrote 1 paragraph.New text: “To opt out of model training, you can turn off "Improve the model for everyone" in ChatGPT under Settings → Data Controls, or select "Do not train on my content" in our privacy portal . For ChatGPT and Codex tasks, either option is sufficie”
View the diff
Privacy PolicyAdded 7 paragraphs and removed 3.New text: “Some U.S. state privacy laws (opens in a new window) require specific disclosures. The following table provides additional information about the categories of Personal Data we collect and how we use and disclose that information. You can ”
View the diff
How your data is used to improve model performanceAdded 4 paragraphs and removed 3.New text: “For a linked teen account, a parent or guardian may manage whether conversations can be used to improve our models through Parental controls .”
View the diff
Sub-processor listAdded 3 paragraphs and removed 2.New text: “Web Hosting: Hosting of ChatGPT Sites created web pages. Applicable subprocessors may run security and safety classifiers on web pages and share results with OpenAI.”
View the diff
How your data is used to improve model performanceRemoved 3 paragraphs and added 2.New text: “One of the most useful and promising features of AI models is that they can improve over time. We continuously improve our models through research breakthroughs as well as exposure to real-world problems and data. When you allow your conten”
View the diff
Privacy PolicyAdded 72 paragraphs and removed 46.New text: “For individuals in the European Economic Area, United Kingdom, and Switzerland, you can read this version of our Privacy Policy.”
View the diff
Sub-processor listAdded 2 paragraphs and removed 1.New text: “ChatGPT Enterprise, Edu & ChatGPT Business :”
View the diff
How your data is used to improve model performanceRewrote 2 paragraphs.New text: “One of the most useful and promising features of AI models is that they can improve over time. We continuously improve our models through research breakthroughs as well as exposure to real-world problems and data. When you share your conten”
View the diff
Privacy PolicyRemoved 29 paragraphs and added 25.New text: “To provide, analyze, and maintain our Services, for example to respond to your questions for ChatGPT;”
View the diff
How your data is used to improve model performanceRewrote 5 paragraphs.New text: “Note : Data retention for certain services may be affected by recent legal developments - please see our blog post for more details.”
View the diff
Privacy PolicyAdded 29 paragraphs and removed 25.New text: “If you live in the European Economic Area (EEA) or Switzerland, OpenAI Ireland Limited, with its registered office at 1st Floor, The Liffey Trust Centre, 117-126 Sheriff Street Upper, Dublin 1, D01 YC43, Ireland, is the controller and is re”
View the diff
Sub-processor listAdded 2 paragraphs and removed 1.New text: “Infrastructure management : Merge API is only used when the Customer chooses to enable Connectors in ChatGPT (opens in a new window) .”
View the diff
How your data is used to improve model performanceAdded 5 paragraphs and removed 4.New text: “When you use our services for individuals such as ChatGPT, Sora, Operator, or Codex we may use your content to train our models.”
View the diff
Privacy PolicyRewrote 2 paragraphs.New text: “We also collect information from other sources, like information that is publicly available on the internet, to develop the models that power our Services. For more information on the sources of information used to develop the models that p”
View the diff
How your data is used to improve model performanceAdded 5 paragraphs and removed 3.New text: “When you use our services for individuals such as ChatGPT, DALL•E, Sora, or Operator, we may use your content to train our models.”
View the diff
Privacy PolicyRewrote 1 paragraph.New text: “To provide, analyze, and maintain our Services, for example to respond to your questions for ChatGPT;”
View the diff
How your data is used to improve model performanceAdded 3 paragraphs and removed 2.New text: “You can also use Temporary Chat from the dropdown in ChatGPT. Chats from Temporary Chat won't appear in history, use or create memories, or be used to train our models.”
View the diff
Privacy PolicyRemoved 74 paragraphs and added 67.New text: “We at OpenAI OpCo, LLC (together with our affiliates, "OpenAI", "we", "our" or "us") respect your privacy and are strongly committed to keeping secure any information we obtain from you or about you. This Privacy Policy describes our practi”
View the diff
How your data is used to improve model performanceRewrote 1 paragraph.New text: “How is data retained in the macOS app?”
View the diff
How your data is used to improve model performanceAdded 9 paragraphs and removed 5.New text: “Learn more about how OpenAI uses content from our services to improve and train our models.”
View the diff
Watched pages
Data controls in the OpenAI platform
OpenAI's documentation on data handling, read live every Monday; 3 answers quote it.
Changed Nov 2, 20265 recordedChecked Nov 9, 2026
How your data is used to improve model performanceOpenAI's help article, read from the Internet Archive's copy of Nov 1, 2026, as the live page refuses automated reads; 3 answers quote it.
Changed Oct 19, 202612 recordedChecked Nov 9, 2026
Privacy PolicyOpenAI's privacy policy, read from the Internet Archive's copy of Nov 8, 2026, as the live page refuses automated reads.
Changed Oct 19, 202611 recordedChecked Nov 9, 2026
Sub-processor listOpenAI's subprocessor list, read from the Internet Archive's copy of Oct 21, 2026, as the live page refuses automated reads; 1 answer quotes it.
Changed Oct 26, 20264 recordedChecked Nov 9, 2026
TermsOpenAI's terms of service, read from the Internet Archive's copy of Feb 26, 2023, as the live page refuses automated reads.
No change seenChecked Nov 9, 2026
About OpenAI
OpenAI (Model APIs, ChatGPT). Answers quote OpenAI's own public pages; a contract, order form or data processing addendum can say more. Company facts live on companies.fru.dev.
Across Glyph Ibis
Compliance: certifications · Companies: Company profile · Acquisitions · awards · blogs · Downtime: 27 incidents in 90 days · Keynotes: Named in keynotes · Leaders: 78 leadership moves · Arena: 20 ranked models · papers · Partnerships: 25 partnerships · Paydays: Paid biweekly, Friday · predictions · Pricing: Model lab · 66 models · Quotes: 50 quotes · Releases: 74 releases · Uptime: 4 services · Events: 9 conferences · Titles: 207 open roles · Training: 3 official courses · Trending: Trending 27 days
Answer history
Trains on app data
on the pagenot on the pagethe quoted clause is no longer on the page; a person re-reads it
Trains on app datanot on the pageon the pagea person re-read the page: the clause is reworded, the answer still holds
Trains on app dataWhen you use our services for individuals such as ChatGPT and Codex, we may use When you use our services for individuals, such as ChatGPT and Codex, we may use your content to train our models.OpenAI rewrote the article between Sep 21 and Sep 23, 2026 (Internet Archive captures of Sep 23 to 26; the live page read in a browser on Sep 27 has the same text). The clause was reworded, not removed; the answer holds and now quotes the new wording.
Trains on API datanot on the pageon the pagea person re-read the page: the clause is reworded, the answer still holds
Trains on API dataBy default, we do not train on any inputs or outputs from our products for businBy default, we don't use inputs or outputs from ChatGPT Business, ChatGPT Enterprise, ChatGPT Edu, or our API to improveOpenAI rewrote the article between Sep 21 and Sep 23, 2026 (Internet Archive captures of Sep 23 to 26; the live page read in a browser on Sep 27 has the same text). The clause was reworded, not removed; the answer holds and now quotes the new wording.
Opt-outnot on the pageon the pagea person re-read the page: the clause is reworded, the answer still holds
Opt-outTo opt out of model training, you can turn off "Improve the model for everyone" To opt out, turn off Improve the model for everyone under Settings > Data controls in ChatGPT, or select Do not train onOpenAI rewrote the article between Sep 21 and Sep 23, 2026 (Internet Archive captures of Sep 23 to 26; the live page read in a browser on Sep 27 has the same text). The clause was reworded, not removed; the answer holds and now quotes the new wording.
Opt-outon the pagenot on the pagethe quoted clause is no longer on the page; a person re-reads it
Trains on app dataon the pagenot on the pagethe quoted clause is no longer on the page; a person re-reads it
Trains on API dataon the pagenot on the pagethe quoted clause is no longer on the page; a person re-reads it