We Use Cookies

    We use cookies to enhance your browsing experience, analyze site traffic, and personalize content. By clicking "Accept All", you consent to our use of cookies. You can customize your preferences or reject non-essential cookies.

    Learn more about our cookie policy
    Model Selection

    OpenRouter Auto (openrouter/auto): AI picks the best model for every prompt

    Stop guessing which model to use. The Auto Router analyzes your prompt and selects the optimal model from Claude, GPT-5, Gemini, Grok, and more — powered by NotDiamond.

    Basics
    ≈ 7 min read
    openrouter/auto

    One Model to Rule Them All — Or Let the Router Decide

    With 300+ models on AI·Collab, choosing the right one for each task can feel overwhelming. Simple questions don't need a frontier model. Complex coding might. The Auto Router (openrouter/auto) solves this: it analyzes your prompt and automatically selects the best model from a curated set of top performers. Powered by NotDiamond's intelligent routing, it considers prompt complexity, task type, and model capabilities. You get great answers without the guesswork — and you only pay for the model that actually runs your request.

    No extra fee — you pay the routed model's rate

    The Auto Router adds no markup. You're billed at the standard rate of whichever model is selected. Check the response to see which model was used.

    AI·Collab also offers free models you can select manually or explore alongside routing. For which models are free, limits, and how that fits your credits, read Free models on AI·Collab (guide) — it complements this Auto Router overview.

    How It Works

    Four simple steps — all invisible to you:

    1

    Prompt analysis

    Your prompt is analyzed by NotDiamond's routing system to understand complexity, task type, and requirements.

    2

    Model selection

    The router picks the optimal model from a curated set — Claude, GPT-5, Gemini, Grok, DeepSeek, Mistral, and more.

    3

    Request forwarding

    Your request is forwarded to the selected model. You get a normal chat experience — the selection happens behind the scenes.

    4

    Response with metadata

    The response includes which model was used. Check OpenRouter Activity or the response model field to see the selection.

    Supported Models (Current Pool)

    The Auto Router selects from a curated set of high-quality models. The pool is updated as new versions ship. As of June 2026, routing includes:

    Anthropic: Claude Haiku 4.5, Claude Sonnet 4.5, Claude Sonnet 4.6, Claude Opus 4.8
    OpenAI: GPT-5, GPT-5 Mini, GPT-5 Nano, GPT-5.1, GPT-5.2, GPT-5.2 Pro, GPT-5.3 Chat, GPT-5.4 Mini, GPT-5.4 Nano, GPT-5.4 Pro, GPT-5.5, gpt-oss-120b
    Google: Gemini 2.5 Flash Lite, Gemini 3 Flash Preview, Gemini 3.1 Flash Lite Preview, Gemini 3.1 Pro Preview, Gemma 4 31B
    DeepSeek: R1, DeepSeek V4 Flash, DeepSeek V4 Pro
    Mistral: Codestral 2508, Mistral Large, Mistral Medium 3.1, Mistral Small 3.2 24B
    xAI: Grok 4.3
    Qwen: Qwen3 235B A22B, Qwen3.6 Plus
    Perplexity: Sonar, Sonar Pro
    Moonshot: Kimi K2 Thinking
    Meta: Llama 3.3 70B Instruct
    MiniMax: MiniMax M2.5
    Inception: Mercury 2
    Z.ai: GLM 5

    Model pool may change. See OpenRouter Auto Router page

    Why Use Auto Router?

    • No more model guessing: Let the router pick the right tool for each prompt
    • Cost optimization: Simpler tasks get routed to efficient models; complex ones get the heavy lifters
    • Quality optimization: Complex prompts automatically go to capable models
    • Great for mixed workloads: Coding, research, quick questions — one model selector handles it all
    • Experimentation: Discover which models work best for your use case without manual A/B testing

    How to Use It on AI·Collab

    Select openrouter/auto from the model dropdown in your chat. That's it. Your prompts will be routed automatically. For reproducible workflows, pin a specific model instead.

    Browse models

    All models — including Auto Router — run on our secure AI platform on German servers. ISO 27001 certified, GDPR compliant, no US access.

    Frequently Asked Questions

    Does Auto Router cost extra?

    No. You pay the standard rate for whichever model is selected. There is no additional fee for using the Auto Router.

    Can the model change between messages?

    Yes. The router selects per request. A simple follow-up might get a fast model; a complex coding task might get Claude Opus or GPT-5.2.

    How do I know which model was used?

    You see it in Usage Statistics under "Model" (Account → Usage Statistics → Model).

    When should I use a fixed model instead?

    Use a fixed model when you need reproducibility — e.g., production workflows, audits, or when you've tuned prompts for a specific model.

    Does streaming work with Auto Router?

    Yes. Streaming and all standard OpenRouter features (tool calling, etc.) work with the selected model.

    Does AI·Collab have free models? How is that related to Auto Router?

    Yes. We maintain a catalog of free models (subject to change). Auto Router picks from the routed model pool for your request; you can also pin a free model manually. See the dedicated guide: Free models on AI·Collab (/blog/free-models).

    What is OpenRouter, in simple terms?

    OpenRouter is an aggregator that gives you access to many AI models — Claude, GPT-5, Gemini, Llama, Mistral and more — through one API and one account, instead of signing up with each provider separately. Its Auto Router (openrouter/auto) goes one step further by automatically choosing the best model for each individual prompt.

    What are the key features of OpenRouter?

    The main features are: (1) a single unified API for 300+ models across providers, (2) automatic model routing (Auto Router) that selects the optimal model per prompt, (3) transparent per-token pricing with no markup on the Auto Router, (4) automatic fallbacks when a provider is unavailable, and (5) a normalized request/response format so you can switch models without rewriting code.

    Is there an EU-hosted, GDPR-compliant alternative to OpenRouter Auto?

    Yes. OpenRouter routes to globally-hosted model endpoints by default, which can complicate EU data-residency requirements. AI·Collab offers the same automatic ‘best model per prompt’ convenience, but with EU-hosted inference, Zero Data Retention, and a GDPR Data Processing Agreement (DPA) — so teams with compliance obligations get smart routing without sending data outside the EU. See our secure platform overview (/sichere-ki-plattform/).

    Related Articles

    Ready to Experience 300+ AI Models?

    Get started today. Access models from OpenAI, Google, Anthropic, Grok and more.

    GDPR compliant · Zero data retention · Cancel anytime