ChatGPT group chat room is now online! Taiwan gets early access to the new 20-person collaboration feature

OpenAI announces new ChatGPT “Group Chat” feature, with Taiwan, Japan, New Zealand, and South Korea included as the first regions to experience it. This feature allows up to 20 people to summon AI in the same chatroom, evolving from a “one-on-one assistant” to a “multi-person collaboration partner.” Free users get 10 AI responses every 5 hours, while Plus subscribers receive 160 responses within 3 hours.

Taiwan Leads the Way in Experiencing the New Era of 20-Person AI Collaboration

(Source: ChatGPT)

OpenAI brings ChatGPT group chat to Taiwan, allowing users to invite up to 20 people to chat together, ushering in the era of multi-person AI collaboration. Taiwan is among the first regions selected, alongside Japan, New Zealand, and South Korea, highlighting OpenAI’s focus on the Asia-Pacific market. This new feature is not just a technical upgrade, but a fundamental change in usage scenarios.

Previously, ChatGPT was mainly a one-on-one personal assistant, with users interacting individually with AI to solve problems. However, in actual work and life, many scenarios require group collaboration: team brainstorming, joint travel planning, co-writing proposals, project direction discussions, etc. Traditionally, each person would ask AI separately and then consolidate the results, which was inefficient and often led to information being out of sync.

The group chatroom feature completely changes this workflow. Now, 20 people can discuss in the same chatroom and summon AI to join the conversation at any time. For example, a marketing team can discuss ad copy together, and whenever they need AI suggestions, simply type “ChatGPT” or “@ChatGPT” for AI to join the discussion and provide input. All members can see AI’s responses, ensuring synchronized information and coherent discussions.

The advantage of this multi-person collaboration model lies in real-time interaction and contextual continuity. AI can read the entire group chat history, understand the flow of the discussion, and provide more contextually relevant suggestions. Compared to the traditional “ask AI separately and consolidate,” the group chatroom greatly reduces communication costs and increases collaboration efficiency.

Step-by-Step Guide to Starting a Group Chatroom

Using group chat is simple. First, click the person icon at the top right of the chat interface, then click “Start group chat” to create a group chatroom. This operation is similar to starting a regular conversation, with an intuitive interface. After clicking, the system automatically generates a new group chatroom and enters the setup page.

Step two: The system provides you with a sharing link, which you can send to the partners you want to chat with. Once they accept, they can join. This sharing link can be sent via any communication tool, including Line, Messenger, Email, or SMS. After clicking the link, the recipient will be directed to the ChatGPT website or app and can confirm to join the chatroom. Members can leave at any time or view the member list, and the management features are quite flexible.

Key Points for Using Group Chatrooms

Create a Group: Top right person icon → Start group chat

Invite Members: Copy the sharing link → Send to partners → They click to join

Summon AI: Type “ChatGPT” or “@ChatGPT” in the chat

View Members: View member list, invite new members, or leave group at any time

Currently, all logged-in accounts can experience this feature, whether free or paid users can create and join group chatrooms. This universality ensures broad application, with no paywall restricting team collaboration.

Differences in Usage Quotas for Free and Plus Plans

If users need AI assistance, simply type “ChatGPT” or “@ChatGPT” in the chat to summon it. This design avoids AI interrupting human conversations repeatedly. AI only responds when explicitly called, ensuring humans remain in control of the chatroom.

Free users get 10 AI responses every 5 hours, while Plus subscribers enjoy 160 responses within 3 hours. This is a significant difference. For free users, you can call AI about once every 30 minutes, suitable for light use or small discussions. The Plus plan allows calling AI more than once per minute on average, supporting intensive collaboration meetings or complex brainstorming sessions.

Importantly, there are no limits on text interactions between members, ensuring discussions are not interrupted by quotas. This design is very user-friendly. The quota only counts AI responses; human-to-human conversation is completely unrestricted. This means teams can discuss freely and only summon AI for expert advice at key moments—fully utilizing AI’s capabilities without affecting discussion flow due to quota limits.

For teams needing frequent AI collaboration, the Plus plan is clearly more valuable. For $20 per month, you get 160 AI call quotas within 3 hours, enabling high-intensity collaboration—very cost-effective for professional teams. The free version is more suitable for occasional AI assistance in casual scenarios, like friends planning trips or discussing interests.

GPT-5.1 Auto’s Smart Context and Politeness Mode

According to the official statement, the new flagship model GPT-5.1 Auto can read group context and automatically switch between GPT-5.1 Instant and GPT-5.1 Thinking based on question complexity. This auto-switching mechanism is a technical breakthrough. Instant mode is for quick responses to simple questions, while Thinking mode handles complex reasoning tasks. The system automatically selects the most appropriate mode based on question complexity, so users don’t need to switch manually.

In group chatroom scenarios, this smart switching is especially important. For simple questions like “What’s the weather today,” Instant mode responds immediately; for complex questions like “How should we plan next quarter’s market strategy,” the system automatically switches to Thinking mode for deeper analysis and reasoning.

The system also has a built-in “politeness mode,” where AI analyzes the conversation flow to determine when to offer suggestions and when to stay silent, reducing concerns about message overload. This is one of the most user-friendly designs of the group chatroom feature. In multi-person conversations, if AI responds immediately every time it’s mentioned, it could interrupt the natural flow, causing information overload.

Politeness mode teaches AI to “read the room.” It analyzes the pace and context of the conversation to judge whether it’s appropriate to provide input. For example, when the discussion is moving quickly, AI may delay its response or simplify its answers; when the discussion hits a bottleneck and needs outside input, AI will actively provide detailed suggestions. This smart judgment makes AI more like a real team member, not just a tool.

Privacy Protection and Safety Mechanisms for Minors

OpenAI states that group chatroom data is completely separated from personal conversations, and AI will not write group content into private “memory.” This privacy design addresses users’ main concerns. Many people use ChatGPT for personal conversations involving private information; if group chat content were mixed into personal memory, it could cause information leaks or improper associations.

This separation mechanism ensures that the group chatroom is an independent space. Content discussed in the group will not affect your personal ChatGPT experience, and AI will not suddenly bring up group chat topics in your private conversations. This isolation protects user privacy and makes the group chatroom a safe space for collaboration.

When minors join, the system automatically lowers the weight of sensitive content. This safety mechanism considers that group chatrooms may involve educational scenarios, such as teacher-student discussions or study groups. When the system detects minor users, it automatically adjusts content filters to reduce the likelihood of inappropriate content, ensuring the safety of minors.