Meta Muse Wonder
Meta's frontier reasoning model for complex personal and professional agentic tasks, with image input, tool calling and a 1M-token context window.
Muse Wonder is a cloud text model from Meta. It is multimodal: alongside a text prompt it accepts images, then replies with generated text. Its context window handles up to 1.05M input tokens and up to 131K output tokens, and it supports adjustable reasoning effort for harder problems. Generation can be tuned with system instructions, temperature, and top-p sampling. It runs through OpenRouter using your own API key, from $4.00 per million input tokens.
- Released
- Oct 2026
- Pricing
- $4.00 / 1M in · $20.00 / 1M out
- Inputs
- Text, Images
- Context window
- 1.05M in · 131K out
- Reasoning
- Adjustable effort
- Controls
- System prompt, Temperature, Top-p
Samples
Generated with Muse Wonder via OpenRouter. The same three prompts run against every text model in the catalog, shown verbatim, so the only thing that changes between two models’ answers is the model.
Explain what an API is to someone who has never written code. Do not use a restaurant or waiter analogy. Plain prose only — no headings, bullet points, or markdown formatting. Maximum 50 words.
An API is a set of rules that lets different computer programs talk to each other and share information. When one program needs something from another, it asks in an agreed format, and the other sends back exactly what was requested.
A two-person startup is deciding whether to build their own user authentication or use a hosted provider. Give them a clear recommendation, then the single strongest argument against your own recommendation. Plain prose only — no headings, bullet points, or markdown formatting. Maximum 150 words.
Use a hosted authentication provider instead of building your own. As a two-person team, your scarcest resource is time, and authentication is undifferentiated work that carries outsized risk, since a single mistake in password storage, session handling, or recovery flows can compromise every user. Providers like Auth0, Clerk, or Supabase Auth give you secure login, multifactor authentication, and social sign in immediately, letting you focus on the product customers actually pay for. The strongest argument against this is vendor dependence, because pricing increases, outages, or a shutdown can force a painful migration and leave you unable to control a critical part of your user experience.
Write a briefing for a non-technical executive explaining why their company's spending on AI is hard to predict, and what they can do about it. Cover at least three distinct causes. Plain prose only — no headings, bullet points, or markdown formatting. Write exactly 205 words.
Your AI bill is hard to forecast because the technology does not behave like traditional software with a fixed monthly license. First, most AI services charge by use, so costs rise and fall with how much your teams and customers ask the system to do, and a popular new feature can multiply usage overnight. Second, the work itself varies widely, since a short question costs less than analyzing a long contract, reviewing thousands of images, or handling a complex conversation that requires several attempts to get right. Third, the underlying technology keeps changing, as providers release new models with different prices, retire older ones, and adjust rates, while your teams experiment and sometimes leave test systems running. You can bring this under control without slowing innovation. Start by tracking spending by team, product, and purpose so you can see what drives each increase. Set monthly budgets and automatic alerts that warn managers before limits are reached, not after. Require a simple cost estimate before launching major features, and review actual results afterward. Choose the smallest, cheapest model that does each job well, reserve premium options for valuable tasks, and shut down unused tools and experiments. Finally, negotiate pricing and review contracts quarterly as usage grows.
Meta
Meta's FAIR and GenAI labs build the open-weight Llama family of language models — among the most widely used open models for local inference.
ai.meta.com ↗Meta's multimodal reasoning model for long-running agentic, multi-agent and coding workflows, with image input and a 1M-token context window.
Meta's open-weight multimodal reasoning model (816B total, 42B active parameters) built for agentic coding, with image input and a 1M-token context window.
Meta Superintelligence Labs' agentic image model, which plans a layout and calls search and coding tools before it renders, for prompt-faithful generation, multi-reference composition and precise local edits.
Meta's speech-to-text model for push-to-talk and speaker-aware transcription, which returns plain text.
Meta's 1B instruction-tuned model — the smallest Llama, runnable in-process via HuggingFace or through Ollama.
Meta's 8B instruction-tuned model with 128K context and strong general reasoning.