It was originally found in the code was leaked and publicized by AI influencers at Xbut OpenAI announced it officially: ChatGPT now offers group chats, enabling multiple users to join the same single ChatGPT conversation and send messages to each other and to the underlying Immense Language Model (LLM), online and via mobile applications.
Imagine adding ChatGPT as another member of your existing group chats, allowing you to text as you would to one of your friends or family members and ask them to respond, and you’ll have an idea of the intriguing power and potential of this feature.
However, for now, this feature is only available in a narrow pilot program for ChatGPT users in Japan, Up-to-date Zealand, South Korea and Taiwan (all tiers including free apply).
“Group chats are just the beginning of ChatGPT becoming a shared space to collaborate and interact with others,” OpenAI wrote in its announcement.
This development is based on internal experimentation at OpenAI, where the technical staff works Keyan Zhang said in a post on X that the OpenAI team initially considered ChatGPT multiplayer to be a “wild idea outside the distribution”.
According to Zhang, the model’s performance in early tests showed much greater potential than existing interfaces typically allow.
The move follows Microsoft, an investor and competitor of OpenAI, updating its Copilot AI assistant last month to enable group chats, as well as Anthropic’s introduction of shared context and chat history from Claude AI models via the Projects feature introduced in summer 2024, although it does not provide real-time, simultaneous group chat in the same way.
Collaboration functionality integrated with ChatGPT
Group chats function as shared conversation spaces where users can plan events, brainstorm, or collaborate on projects with additional ChatGPT support.
These conversations are distinct from individual conversations and are excluded from ChatGPT’s memory system – meaning that no data from group threads is used to train or personalize future interactions.
Users can start a group chat by selecting the people icon in a up-to-date or existing conversation. Adding others creates a copy of the original thread while retaining the source dialogue. Participants can join using a shared link and will be asked to create a profile with their name, username and photo. This feature supports 1 to 20 participants per group.
Each group chat is listed in a up-to-date section of the ChatGPT interface, and users can manage settings such as naming the group, adding or removing participants, and muting notifications.
Supported by GPT-5.1 with extended tools
The up-to-date group chat feature runs on GPT-5.1 Auto, a backend setting that selects the optimal model based on the user’s subscription level and prompt.
Features such as search, image generation, file transfer and dictation are available in group conversations.
Importantly, the system applies bid limits only when ChatGPT generates responses. Direct messages between users in a group do not count against any plan’s message limit.
OpenAI has added up-to-date social features to ChatGPT to support the dynamics of this group. The model can respond with emoticons, interpret the context of the conversation to decide when to respond, and personalize the generated content using members’ profile photos, for example by inserting users’ likenesses into the images when asked to do so.
Privacy by default, control for younger users
OpenAI emphasized that privacy and user control are integral to the group chat design. This feature works independently of the user’s personalized ChatGPT memory and no up-to-date memories are created as a result of these interactions.
Participation requires an invite link, and members can always see who’s in the chat or leave the chat at any time.
Users under 18 are automatically protected from sensitive content in group chats. Parents or guardians can completely disable group chat access using the built-in parental controls.
Group creators retain special privileges, including immunity from deletion by others. All other participants can be added or removed by group members.
A testbed for collaborative AI experiences
OpenAI sees group chats as an early step towards richer multi-user AI applications, pointing to broader ambitions for ChatGPT as a shared workspace. The company expects to expand access and improve this feature over time based on how early users apply it.
Keyan Zhang’s post suggests that the capabilities underlying the model far outweigh the interfaces that users currently interact with. According to OpenAI, this pilot offers a up-to-date “container” in which more of the model’s hidden capacity can be revealed.
“Our models have much more room to shine than today’s experience shows, and current containers only use a fraction of their capabilities,” Zhang said.
Since this initial pilot is focused on a narrow number of markets, OpenAI is likely monitoring both usage patterns and cultural fit as it plans for a broader rollout. For now, the group chat experiment offers users a up-to-date way to interact with ChatGPT – and with each other – in real time, using a conversational interface that combines productivity and personalization.
Developer Access: Still unclear
OpenAI has not provided any information that group chats will be available via API or SDK. The current implementation relies solely on the ChatGPT product framework, with no mention of tool calls, developer hooks, or integration support for programmatic apply. This lack of signaling leaves it unclear whether the company views group interaction as a primitive future developer or as a contained UX feature solely for end users.
For enterprise teams exploring ways to replicate multi-user collaboration using generative models, any current implementation would require custom orchestration – such as managing multi-party context and prompts in separate API calls, and supporting linking session state and responses externally. Until OpenAI provides formal support, group chats remain a closed-interface feature rather than a developer-accessible feature.
Here is a standalone final subsection adapted to the article, focusing on the importance of implementing ChatGPT group chat for enterprise decision makers both in pilot regions and around the world:
Implications for AI enterprises and data leaders
For enterprise teams already using or preparing to apply AI platforms, OpenAI’s group chat feature introduces a up-to-date layer of multi-user collaboration that could transform the way generative models are implemented in workflows. While the pilot program is narrow to users in Japan, Up-to-date Zealand, South Korea and Taiwan, its design and roadmap provide key signals to AI engineers, orchestration specialists and data leaders around the world.
AI engineers managing gigantic language model (LLM) implementations can now begin to conceptualize real-time, multi-user interfaces not only as support tools, but also as collaborative environments for research, content generation, and ideation. This adds another front in model tuning: not only how models respond to individuals, but also how they behave in live group settings with changes in context and different user intentions.
For AI orchestration leaders, the ability to integrate ChatGPT into collaboration flows without exposing private storage or having to create custom builds can reduce friction in piloting generative AI in cross-functional teams. These breakout sessions can serve as a lightweight alternative to internal brainstorming, prototyping or knowledge sharing tools – useful for teams narrow by infrastructure, budget or time.
Enterprise data managers can also find apply cases in organized group chat sessions for data annotation, taxonomy validation, or internal training support. The lack of memory persistence in the system provides a level of data isolation consistent with standard security and compliance practices – although global implementation will be key to validating regional data handling standards.
As group chat capabilities evolve, decision makers should monitor how shared usage patterns may impact future model behavior, audit needs, and governance structures. In the long term, these types of features will impact not only how organizations interact with generative AI, but also how team-level interfaces are designed around it.