KI-Modelle erkunden

Anthropic's most capable model available. Built for the hardest coding, research and long agent tasks, with a million-token context. Use it when quality matters more than cost.

Google's lowest-cost Gemini. Good for quick replies, simple questions and high-volume chat where speed and price matter most.

Very fast and very affordable. Good for quick lookups, summaries and simple everyday questions.

xAI's coding model, built for agent-style software work. It always reasons before answering, reads screenshots and diagrams, and handles refactors, debugging and multi-step coding tasks at a low price.

Tencent's model built for real business work. It handles coding, long documents and multi-turn chat well, and balances reliability, speed and cost for both quick replies and complex workflows.

Answers backed by live web search. Great for staying current, quick research, and fact-checking in real time.

Anthropic's newest Opus model, at a lower price than Opus 5. Excellent at hard coding problems, deep analysis and high-quality writing, and it always thinks before answering. Has a million-token context.

A smart and fast model that handles harder questions while staying quick and affordable for everyday use.

A fast, budget-friendly model for everyday writing, answering questions and light analysis.

A capable Grok model with built-in reasoning, live web search and a million-token context. Good for deep analysis and research-heavy chats at a lower price than the newest Grok models.

Meta's model built for long, detailed conversations. It can read images and handle large documents, which makes it good for research and analysis.

Anthropic's previous Opus model, built for complex, demanding work. Excellent at hard coding problems, deep analysis and high-quality writing, with a million-token context.

A strong reasoning model from the Gemini 2.5 series. Good at math, tough coding problems and deep analysis, and it handles very long documents well.

A smart, fast model with web search built in. Good for quick tasks where you also want up-to-date information.

A strong Grok model for coding, agent tasks and knowledge work. It has built-in reasoning and live web search, which makes it good for deep analysis and research-heavy chats.

A strong Meta model that punches above its price. Good quality for coding, analysis and writing at a fraction of the usual cost.

A fast, capable Claude model that gets close to Opus quality on coding and agent work at a much lower price. A strong everyday pick for coding, analysis and writing, with a million-token context.

Google's open model. It understands images, writes code and works in 140+ languages, which makes it a good pick for everyday tasks.

A well-rounded model that balances speed and quality. Solid for coding, writing and tasks that need a bit more depth.

A powerful Grok model built for long-running coding and agent work that needs little hand-holding, plus questions that need real-world knowledge. It reads images and supports live web search.

A compact model that follows instructions well and works across many languages. A good fit for international use.

An earlier Opus model from Anthropic. Still very strong at difficult questions, high-quality writing and complex coding.

A low-cost Gemini 3 model with web search and sharp instruction following. Good for frequent, everyday conversations.

A capable mini model with web search. Strong at coding and good for detailed, research-backed conversations.

Grok 4.7 is xAI's latest flagship. It checks its own work more carefully and stays on track through long, multi-step tasks, which makes it a strong pick for coding, research and in-depth knowledge work. It reads images and supports live web search.

Alibaba's multimodal MoE model and an early preview of the Qwen4 architecture. It reads text, images and video across a million-token context, and is built for coding and agent work at a fraction of frontier model cost.

The fastest Claude model on Aymo. Great for quick questions, short tasks and back-and-forth conversations where speed matters, with optional step-by-step thinking.

The balanced middle model of the GPT-5.6 family. Fast, reliable and a strong everyday pick for coding, analysis and writing, with a million-token context.

A capable Gemini 3.5 model with web search and image understanding. Strong at coding and longer, more complex chats.

A fast, low-cost Gemini made for simple jobs you run over and over, like search, translation and document processing. It keeps the same million token context as the bigger models, so long inputs still work.

Google's Flash 3.6 model is a solid default for most work. It reads text, images, video, audio and PDFs, holds a million tokens of context, and was built for coding and multi-step agent tasks.

Google's open model. It understands images, writes code and works in 140+ languages, which makes it a good pick for everyday tasks.

The lowest-cost model in the GPT-5.6 family. Quick and capable for high-volume everyday tasks, with web search and a million-token context.

A fast Gemini 3 model built for coding and agent work. It handles text, images, audio, video and long documents, and supports web search.

Gemini 3.8 Flash is Google's smartest Flash model yet. It's a clear step up from 3.7 Flash in coding, agent tasks and multi-step reasoning, while staying fast. It works with images, audio, video and long documents, and supports web search.

OpenAI's open-weight reasoning model. Strong for its price at coding, math and analysis, and one of the cheapest capable models available.

A powerful OpenAI model on Aymo. Excellent at complex coding, in-depth analysis and professional-grade writing.

A reasoning model that shows its thinking as it works. Good for hard coding problems and thorough analysis.

Google's most advanced model on Aymo. Best for in-depth reasoning, long documents, and complex tasks that need the full power of Gemini.

A strong OpenAI model for complex coding, professional writing and deep analysis. Cheaper than GPT-5.5 while keeping the quality high.

OpenAI's earlier GPT-5 flagship. Still very good at coding, writing and agent tasks, at a lower price than the newest models.

Kimi K3 is Moonshot AI's latest open-weight flagship. It brings frontier-level coding, deep reasoning, native vision, a million-token context and strong knowledge work performance.

The top model of the GPT-5.6 family. Great for agent work, long multi-step tasks and hard coding problems, with a million-token context.

Moonshot's coding model, built on Kimi K2.6 and tuned for long software projects. It plans, writes, tests and debugs code across many steps, and uses fewer thinking tokens than K2.6.

The capable workhorse of the GPT-6 family. Strong at coding, reviewing code, debugging and analyzing data, at half the price of GPT-5.6 Sol. Supports web search.

A multimodal agent model built for long coding sessions. It reads images, turns simple prompts into working interfaces, and stays reliable across long, complex tasks.

The lowest-cost model in the GPT-6 family. Quick and capable for summaries, pulling out key details and straightforward questions, with web search.

A dependable model with strong coding skills and image understanding. It holds up well across long, involved conversations.

DeepSeek's newest model and the first on its new architecture. It beats DeepSeek's own V4 Pro on coding and agent tasks at a much lower price, with a million-token context.

DeepSeek's V4 flagship and one of the strongest open models available. A top performer for coding, deep analysis and complex problem solving.

A well-balanced DeepSeek model that is good at coding and reasoning without breaking the bank. A solid value pick.

A smart reasoning model that gets close to top-tier results at a much lower cost. Great value for analysis and coding.

A multimodal reasoning model that reads text, images and video across a million-token context. Built for long, complex conversations and demanding coding and analysis work.

Z.ai's latest flagship, with big gains in complex software engineering and agent work. It uses the same base model as GLM-5.2 and supports a million-token context.

A flagship model built for long-horizon tasks, with a million-token context that holds up in practice. Great for long, autonomous work sessions and complex development.

A highly ranked model for coding and building. Strong at long, autonomous work sessions and complex development.

A fast multimodal model from Z.ai that reads text, images and video. Built for efficient coding and long agent tasks, with an architecture that keeps long-context answers accurate while using less compute.

A fast, responsive model that balances quality and speed. A solid pick for frequent, cost-conscious use on Aymo.

A 27B open vision-language model with native multimodal support and a million-token context. Strong for its size at coding and analysis.

Qwen's flagship 2.4T MoE model with native multimodal input and a million-token context. Great for large-scale code, planning and knowledge work.

AI Image Generator. Output resolutions: 1K, 2K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K, 2K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K, 2K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 2K, 4K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 2K, 4K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K, 2K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K, 2K, 4K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K, 2K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K, 2K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K, 2K, 4K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.

AI Image Generator. Output resolutions: 1K, 2K, 4K. Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16.