Poe just added multi-model group chats for up to 200 people

For the past two years, interacting with AI chatbots has felt like being in a strictly one-on-one direct message. Quora's Poe is ready to turn that room into a crowded party.

On Monday, Poe launched a new group chat feature worldwide, letting up to 200 people jump into a single conversation alongside more than 200 different AI models covering text, image, video, and audio generation.

The launch lands just days after OpenAI started piloting its own group chats for ChatGPT across select markets including Japan, South Korea, New Zealand, and Taiwan. But while OpenAI is testing the waters slowly with its own tool, Poe is opening the floodgates globally with an entire roster of rival bots.

Bringing the entire AI ecosystem into one group DM

Users can start a group chat right from the Poe homepage, with real-time syncing between desktop and mobile so conversations don't break when switching devices.

Quora spent six months building this setup, and the use cases they imagine are surprisingly everyday. Friends planning a vacation can tap Gemini 2.5 for search while simultaneously querying o3 Deep Research for detailed travel itineraries. Design teams can brainstorm mood boards across multiple image models right in front of each other, or friends can hang out and play trivia games using creator-built quiz bots.

The model lineup available within these group threads is basically a who's who of current frontier tech. Users can ping Claude 4.5 Sonnet, GPT-5.1, or o3 Deep Research, call up ElevenLabs v3 and ElevenLabs Music for audio, or generate video clips with Sora 2 Pro, Veo 3.1, and Kling 2.5 Turbo Pro without ever leaving the chat.

Quora also made sure creator-made custom bots, like Nano Banana or tailored persona bots, can be brought into group chats, leaning heavily into their community library.

Why it matters

We think this launch highlights the next big battleground for consumer AI: turning solitary chatbots into multi-player collaborative platforms.

For remote teams, creative groups, and everyday friends, having multiple AI models in the same room eliminates the annoying routine of copying and pasting outputs across Slack or WhatsApp. You can literally ask one model to draft a concept and immediately prompt another to critique it or turn it into video, all in plain view of your teammates.

It is also a brilliantly defensive move against OpenAI. ChatGPT might dominate consumer mindshare, but it is locked into OpenAI's own walls. Poe's single biggest asset has always been aggregation, and letting users pit Claude against GPT and Sora against Veo in the same live conversation plays directly to that strength.

The big question mark, of course, is conversation clutter and credit burn. When 200 humans and several heavy-duty models share a thread, context windows and point quotas could drain fast. But as a proof of concept for multiplayer AI, Poe just threw down a serious gauntlet.