OpenAI Mixpanel Breach: What User Data Was Exposed?

The Mixpanel data breach affected some users of OpenAI’s API platform: information in analytics records associated with their API accounts was exposed, not their API conversation content or credentials. The exposed fields could include a user’s name, email address, approximate location inferred from a browser, operating system and browser details, and referring websites or campaign information.

What happened

OpenAI said it was notified on November 9, 2025, that an attacker had gained unauthorized access to part of Mixpanel’s systems and exported a dataset containing limited analytics information about some OpenAI API users. Mixpanel provides product analytics: services that help companies understand how people use a website or application. OpenAI used it to analyze activity on the API platform.

The incident was in Mixpanel’s environment, rather than an intrusion into OpenAI’s own systems. OpenAI said it ended its use of Mixpanel, investigated the affected records, and notified impacted organizations and users. The scope was limited to certain API users; the public notice did not say that every OpenAI account was involved.

What information was exposed

OpenAI described the exported records as limited profile and analytics information connected to use of the API platform. Depending on the person’s record, exposed data could include:

• Name and email address associated with an API account.
• Approximate location based on a browser’s general region, such as city, state, or country.
• Browser and operating-system information, including the browser or device environment used to visit the platform.
• Referring websites and campaign information, such as the source that led a user to an OpenAI page.
• Organization or user identifiers associated with the API account.

These details should not be treated as a complete list of fields for every affected person. Analytics records vary, and the incident notice describes what might have appeared in the exported data, not that every field was present for every user.

What was not exposed

OpenAI said the incident did not expose chat conversations, API requests or API usage data, passwords, API keys, payment details, or government-issued identification. It also said the incident did not involve a breach of OpenAI’s systems. This distinction matters: the reported data was about account identity and how someone reached or accessed the API platform, rather than the content they sent through the API.

That does not make the exposed information harmless. A name and email address, paired with knowledge that someone uses the API platform, can make a phishing message sound credible. But a stolen analytics record is not itself an API key or password, and OpenAI did not say the attacker could use the exposed fields to log in to accounts.

Who may be affected

The affected group was some users of OpenAI’s API platform. The notice did not describe the incident as affecting all ChatGPT users, and it did not say that ChatGPT conversations were included. People who only use ChatGPT should not assume they were affected based on the Mixpanel incident alone. Conversely, API users should check for a direct notification rather than trying to infer their status from the general description.

OpenAI said it would contact affected users and organizations. If you received a notice, use the account or organization contact channels you normally trust to verify its origin. Avoid clicking a link in an unexpected email; instead, open the relevant service directly or ask your organization’s administrator. This is a sensible precaution for any breach notification, not evidence that a particular message is fraudulent.

What affected users should do

OpenAI said API keys and passwords were not part of the exposed dataset, so it did not identify a need to rotate credentials solely because of this incident. Still, users and organizations can take practical steps to reduce the risk of impersonation:

1. Read any direct notice carefully and confirm whether it says your account or organization was affected.
2. Be cautious with messages that claim to be from OpenAI or a related service, especially if they ask for a password, API key, payment, or urgent action.
3. Do not share API keys in response to an email or support request. Use the platform’s normal account tools to manage credentials if you have a separate reason to change them.
4. Tell colleagues who handle API access or billing that contact details may have been exposed, so they can recognize targeted impersonation attempts.
5. Report suspicious messages through the service’s official support or security channels, reached independently rather than through a message link.

If you did not receive a notice, that alone is not proof of either inclusion or exclusion; follow any updates from OpenAI or your organization. Do not use unofficial “breach lookup” pages that ask for account credentials.

Why analytics data can be involved

Product analytics tools record events that help a company measure how a service is used—for example, which page a visitor opened, what kind of browser they used, or which referral brought them there. A record may include an account identifier so activity can be connected to a user or organization. Such tools are distinct from the systems that store message content or authenticate logins, although the data they hold can still identify people.

The breach also illustrates that a company’s security depends in part on the vendors it uses. A service provider may process information on behalf of a customer, and an incident at the provider can expose that information without a direct compromise of the customer’s own network. OpenAI’s response included ending its use of Mixpanel and reviewing the affected data and its vendor relationships.

Frequently Asked Questions

Were ChatGPT conversations exposed in the Mixpanel breach?
No. OpenAI said chat conversations and API request content were not exposed in this incident.

Were OpenAI API keys or passwords stolen?
OpenAI said API keys and passwords were not included in the exported Mixpanel data.

Did the breach affect every OpenAI user?
No. OpenAI described the affected population as some users of its API platform. It did not say all API users or all ChatGPT users were affected.

What personal details could have been exposed?
Depending on the record, information could include a name, email, organization or user identifier, approximate location, browser and operating-system details, and referral or campaign information.

Should I change my API key because of this breach?
OpenAI said API keys were not exposed, so the incident alone was not identified as a reason to rotate one. Follow any direct guidance from OpenAI or your organization, and never send a key to someone who contacts you unexpectedly.

How can I tell if I was affected?
Check for a direct notification from OpenAI or your organization’s administrator. The public description does not provide a way to determine an individual’s status from the exposed fields alone.

Related

• Clean Your Mac: A Complete Guide
• Basic Array Formula Example: Excel & Google Sheets
• Measure Noise Levels Apple Watch: How-To