Poe
Chat platform that bundles multiple mainstream AI models, supporting writing, coding, image generation, and web search
About Poe
Poe is a chat platform that aggregates multiple large models. The official Explore page explains that questions are automatically routed by task and subscription status, so writing, coding, real-time information, and image generation can be completed directly. The platform hosts models such as GPT-6 Astra, Claude Sonnet 5.5, Gemini Omni 1.1 Flash, and Mistral Large 4, and supports 1M token context windows, native vision, and web search for complex reasoning, agentic coding, and document creation. Gemini Omni 1.1 Flash can generate video from text, images, or video, extending scenes or interpolating between the first and last frames. Client downloads for iOS, Android, macOS, and Windows are also available.
Best for
Best use cases
Everyday writing and drafting
Programming Q&A and agentic code tasks
Image generation and image understanding
Long-document analysis and web research
Features
Multi-model aggregation and automatic routing
The Explore page lists models such as GPT-6 Astra, Claude Sonnet 5.5, Gemini Omni 1.1 Flash, and Mistral Large 4, routing each question to the matching model by task and subscription status.
Image and video generation
Images can be generated directly; Gemini Omni 1.1 Flash generates video from text, images, or video, extending scenes or interpolating between the first and last frames for coherent camera motion.
Long context and native vision
GPT-6 Astra and Claude Sonnet 5.5 provide 1M token context windows, native vision, adjustable reasoning effort, and web search, handling screenshots and documents.
Clients on multiple platforms
The official Get Poe page provides client download entries for iOS, Android, macOS, and Windows.
Pros
- Aggregates mainstream models like GPT, Claude, Gemini, and Mistral in a single interface
- Routes questions to a suitable model by task and subscription status, reducing model-selection overhead
- GPT-6 Astra and Claude Sonnet 5.5 support 1M token context windows and native vision for complex documents
- Official client downloads for iOS, Android, macOS, and Windows
Cons
- Mistral Large 4 is in public preview (officially labeled Available in public preview)