OpenAI is slowing down the release of ChatGPT’s new voice mode, shifting the first small test from late June to late July. The company says the delay is tied to safety work, product refinement, and infrastructure readiness before the feature reaches all ChatGPT Plus users in the fall.
What changed in the rollout plan
The original plan was to begin the rollout in late June with a small test group of ChatGPT Plus subscribers. That schedule has now moved back by a month.
OpenAI now expects the rollout to begin in late July. The release will be iterative, which means access will expand in stages rather than arriving for every Plus subscriber at the same time.
The company’s current plan is for all Plus users to gain access in the fall. That timeline is not fixed, because OpenAI says the exact schedule depends on whether the system meets security and reliability standards.
Why OpenAI is delaying the feature
The delay is being framed as a safety decision. OpenAI says it is still refining the system before allowing broader use of ChatGPT’s advanced voice capabilities.
A key part of that work is content moderation. The company says it is improving how the system rejects inappropriate material, which suggests the voice feature needs to handle difficult or unsafe requests reliably before it can scale to more users.
OpenAI is also working on the user experience and the underlying infrastructure. Those two areas matter for different reasons: the user experience shapes how people interact with the voice mode, while infrastructure affects whether the service can operate dependably as access expands.
- Safety: OpenAI says the launch moved back for safety reasons.
- Moderation: The company is improving content moderation and rejection of inappropriate material.
- User experience: OpenAI says it is continuing to refine how the system works for users.
- Infrastructure: The company is also improving the systems behind the feature.
- Reliability: Full access depends on meeting reliability standards.
What Plus users should expect
ChatGPT Plus subscribers should not expect an immediate full release. The first step is now a limited rollout beginning in late July, aimed at a small test group.
That staged approach gives OpenAI room to evaluate the system before expanding availability. The source does not describe how users will be selected for the first group, so the practical takeaway is simple: Plus access is planned, but not guaranteed on a single launch day.
The broader target remains the fall. However, OpenAI has tied the final timing to security and reliability standards, so the schedule could depend on how the late July rollout performs.
For users waiting on ChatGPT voice mode, the important distinction is between the start of testing and full availability. Late July marks the beginning of the iterative rollout, while fall is the planned window for all Plus users.
Video and screen sharing are separate
OpenAI is also developing new video and screen sharing features. Those features are not part of the same release schedule as the advanced voice capabilities.
The company says video and screen sharing will be released separately. That means users should not assume that voice, video, and screen sharing will arrive together as one combined update.
This separation matters because each feature likely has its own readiness requirements. Based only on the source, OpenAI is presenting voice mode as the feature now moving into an iterative rollout, while video and screen sharing remain separate developments.
The bigger product signal
The delay shows that OpenAI is prioritizing a controlled launch over a faster full rollout. The company is not canceling ChatGPT’s advanced voice capabilities, but it is adding time before the feature reaches every Plus subscriber.
That choice puts security, reliability, content moderation, user experience, and infrastructure at the center of the release. For a voice feature, those factors are not secondary details. They shape whether the product can respond appropriately, run consistently, and support wider access.
The current schedule is therefore best understood as conditional. A late July test is the next planned step, and fall access for all Plus users remains the goal. But OpenAI has made clear that the timeline depends on whether the system meets the standards it has set for safety and dependable operation.