Poe

OpenAI's GPT-5.4-Nano and GPT-5.4-Mini are now live on Poe. GPT-5.4-Nano is a strong fit for fast, high-volume tasks like summarizing transcripts, labeling tickets, rewriting content, quick RAG answers, and running @openclaw flows with Poe where latency and cost matter most. GPT-5.4-Mini is better for tasks like turning messy emails or notes into clean JSON, fixing a broken function without rewriting the whole file, deciding how to respond to an unusual support ticket, or carrying out an agent task that needs to plan, check results, and take a next step. Try them today alongside every top model in the Poe app and API today at poe.com/GPT-5.4-Nano and poe.com/GPT-5.4-Mini.

MiniMax-M2.7 is live on Poe! A next‑gen self‑evolving model built for autonomous software engineering and agent workflows. M2.7 can iteratively improve its own agent scaffolds, optimize task performance over repeated runs, and deliver major gains on real-world coding benchmarks like SWE-Pro. Great for multi-file codebase edits, long-horizon planning, tool use, document automation, and complex multi-agent tasks. You can try it in Poe app on all platforms and in the Poe API at poe.com/Minimax-M2.7

Mistral Small 4 is live on Poe! One model for reasoning, vision, and agentic coding. Good for things like solving multi-step math or research problems, extracting structured data from images or scanned documents, debugging and navigating a codebase, analyzing a chart and summarizing the takeaway, or parsing a dense PDF and pulling out what matters. You can try it in Poe app on all platforms and in the Poe API at poe.com/Mistral-Small-4

Grok 4.20 Multi-Agent is live on Poe! Instead of one model doing everything sequentially, @xai's new Grok 4.20 Multi-Agent spins up multiple specialized agents that search, analyze, and synthesize in parallel. Ask it your hardest questions and watch as it spawns multiple agents in parallel server-side to produce a rigorously sourced answer with citations in minutes. You can try it in the Poe app on all platforms and in the Poe API at poe.com/Grok-4.20-Mult…

We had early access to GPT-5.4 and GPT-5.4 Pro, and they’re now available on Poe. In our testing, the models were particularly strong at long workflows: debugging code, chaining tools together, and running deep research across many sources. The big difference we noticed: persistence. They keep iterating on hard problems where earlier models often stop. We also ran our own BrowseComp-style research eval and saw the same pattern—they were much better at “needle-in-a-haystack” queries that require digging across many pages and synthesizing the answer. Combined with much longer context, they can track complex tasks without losing the thread. GPT-5.4 is available at poe.com/GPT-5.4 and GPT-5.4-Pro at poe.com/GPT-5.4-Pro. Try them across all Poe platforms or via the Poe API.

Try it alongside every top model in the Poe app and API today at poe.com/GPT-5.3-Codex-…. Pricing is 10 points/message to ensure broad access.

GPT-5.3-Codex-Spark is now available on Poe in a limited research preview! Poe is one of the first places where this model is available today. Enabled by Cerebras’s high-performance chips, OpenAI's newest lightweight coding model is optimized for ultra-fast code generation (~1000s tok/s) and supports a 128k context window.

GPT-5.3-Instant is live on Poe! OpenAI's fastest GPT-5.3 model delivers near-instant responses while retaining strong reasoning and instruction-following. Built for high-throughput workflows where speed matters as much as quality. You can try it in Poe app on all platforms and in the Poe API at poe.com/GPT-5.3-Instant

Gemini 3.1 Flash-Lite is live on Poe! Gemini 3.1 Flash Lite is fast, cost-efficient lightweight reasoning model built for ultra-low latency workloads. Supports web search and up to 1 million tokens of input context, built for fast, scalable performance across real-time applications. You can try it in Poe app on all platforms and in the Poe API at poe.com/Gemini-3.1-Fla…

Kling 3.0 family is live on Poe! Kling 3.0 is a next-generation cinematic video model capable of creating up to 15-second narrative videos with multi-shot storyboarding, consistent characters and voices, physics-driven camera motion, and native audio. Kling O3 adds real-time optimization for faster, cost-efficient generation.  You can try them in Poe app on all platforms and in the Poe API at poe.com/Kling-v3, poe.com/Kling-v3-Pro, poe.com/Kling-O3

Seed 2.0 mini is live on Poe! ByteDance's latest model supports 256k context, image and video understanding, native web search, and four reasoning modes ranging from fast to deep. A strong option for developers who need flexible, cost-efficient reasoning. You can try it in Poe app on all platforms and in the Poe API at poe.com/Seed-2.0-Mini

Seedream 5.0 Lite is live on Poe! A high-fidelity image generation model that integrates multi-step visual reasoning and precise control. Built with a unified multimodal architecture, it delivers superior character consistency, advanced editing capabilities, and enhanced world knowledge for production-level reliability across creative workflows. You can try it in Poe app on all platforms and in the Poe API at poe.com/Seedream-5.0-L…

Qwen3.5 Flash is live on Poe! A fast and efficient multimodal model that processes text and images seamlessly. Built with linear attention and a sparse mixture-of-experts architecture, it delivers improved inference efficiency and stronger performance than the 3 series across both text and vision tasks. You can try it in Poe app on all platforms and in the Poe API at poe.com/Qwen3.5-Flash

Nano Banana 2 is live on Poe! Google’s latest image model combines Pro-level intelligence with lightning-fast generation. It features advanced world knowledge, precise text rendering, strong subject consistency, and 4K production-ready outputs. You can try it in Poe app on all platforms and in the Poe API at poe.com/Nano-Banana-2

GPT-5.3-Codex is live on Poe! OpenAI's most capable agentic coding model combines advanced reasoning with autonomous coding workflows. 25% faster than its predecessor, built for tasks you can start and walk away from. Try it alongside every top model in the Poe app and API at poe.com/GPT-5.3-Codex

@DnuLkjkjh No user data was used in the making of the custom benchmark, to be clear!

@poe_platform "Co-developed a custom benchmark" is a very creative way to describe being a test environment. Smart positioning though — attaching yourself to the release narrative is worth more than the benchmark itself.

Gemini 3.1 Pro is live on Poe — Google’s newest Gemini model built to solve your hardest challenges: deep reasoning, synthesis, and complex problem solving. We got early access to a private snapshot and co-developed a custom benchmark to help shape this model with our partners at Google DeepMind. Try it now in the Poe app on all platforms and in the Poe API: poe.com/Gemini-3.1-Pro

Claude Sonnet 4.6 is Anthropic’s latest model. It strikes a strong balance of quality, speed, and price for day-to-day work, with robust computer-use capabilities for tool-driven workflows and UI automation. You can try the model in the Poe app on all platforms and in the Poe API at: poe.com/Claude-Sonnet-…

Qwen3.5-Plus is live on Poe! Qwen3.5‑Plus is a native vision‑language model built on a hybrid linear‑attention and sparse MoE architecture for greater inference efficiency. It features a 1M‑token context window, enabling powerful long‑context multimodal reasoning and agent workflows. Compared to Qwen3, it represents a major leap in both pure‑text and multimodal capabilities, performing on par with leading state‑of‑the‑art models. You can try them in the Poe app on all platforms and in the Poe API at: poe.com/Qwen3.5-Plus