Debuly
Meta Model API logo

Meta Model API

Pay-as-you-go API for Meta's Muse models covering text, image generation, speech transcription and segmentation, OpenAI SDK compatible.

Paid

What is Meta Model API?

Meta Model API is a self-serve developer API for Meta's Muse family of models. Text models Muse Wonder and Muse Spark handle agentic work, coding, reasoning, tool calling, structured output and search grounding with text, image, video, audio and PDF input and a context window of about one million tokens. Other endpoints cover Muse Image for image generation and editing, Muse Voice Transcribe for speech to text, and Segment Anything Model (SAM) 3.1 for image and video segmentation. You create an API key, point an OpenAI SDK compatible client at it, and call the Responses, Chat Completions or Messages endpoints. Billing is pay-as-you-go with no minimums: Muse Spark costs $1.25 per 1M input tokens and $4.25 per 1M output tokens, Muse Image is $0.01 per image, and a discounted Contributor mode trades lower prices for permission to train on your content. Meta also offers Muse Code for terminal coding.

Use cases

  • Build agents that call tools in parallel
  • Generate and review code across repositories
  • Ground answers with live web search
  • Generate images through an API
  • Transcribe audio to text
  • Segment objects in images and video
  • Prototype cheaply in Contributor mode

Who it's for

  • Developers
  • AI engineers
  • Startups building agents
  • Research teams

Why use Meta Model API

  • Billing is per use with no minimums or upfront commitment, so small experiments stay cheap.
  • Existing OpenAI SDK code can be pointed at the API, which shortens migration.
  • There is no long-context premium, so full-context requests cost the same per token.
  • One account covers text, image, speech and segmentation models.
  • Standard mode does not use your content to train Meta models, and Contributor mode is optional.

Meta Model API features

  • Muse Wonder and Muse Spark models with a 1M-token context window
  • Input of text, images, video, audio and PDFs
  • Responses, Chat Completions and Messages API endpoints
  • Reasoning effort levels, tool calling and structured output
  • Web search grounding at $2.50 per 1,000 queries
  • Prompt caching with a lower cached-input rate
  • Muse Image generation at a flat $0.01 per image
  • Muse Voice Transcribe for streaming and file speech to text
  • Segment Anything Model 3.1 for image and video segmentation
  • Standard and Contributor modes with different data-use terms

Meta Model API integrations

  • OpenAI SDK
  • OpenRouter
  • Muse Code

How to get started with Meta Model API

  1. Sign in at dev.meta.ai and generate an API key.
  2. Point an OpenAI SDK compatible client at the Meta Model API base URL.
  3. Choose a model such as muse-spark-1.3 or muse-wonder-1.0 and call the Responses or Chat Completions endpoint.
  4. Add tools, web search grounding or structured output as needed and watch usage in the response.
  5. Monitor rate-limit headers and retry 429 responses with exponential backoff.

Meta Model API limitations

  • Muse Wonder is in preview and is not available in South Korea
  • Contributor mode lets Meta train on your prompts, code and responses
  • Contributor-mode Muse Spark variants have lower limits of 150 requests per minute
  • Audio understanding in Muse Spark 1.3 may be degraded, so use 1.2 or Muse Voice Transcribe
  • Model availability can vary by country

Meta Model API pricing

Meta Model API has no free plan.

Meta Model API plans

  • Muse Spark

    $1.25 input /$4.25 output per 1M tokens

    pay-as-you-go

    • Cached input $0.15 per 1M tokens
    • Contributor mode $0.10 input, $0.20 output
  • Muse Wonder

    $4.00 input /$20.00 output per 1M tokens

    pay-as-you-go

    • Cached input $0.20 per 1M tokens
    • Standard mode only on Model API
  • Muse Image

    $0.01 per image

    pay-as-you-go

    • Built-in web and image search included
  • Muse Voice Transcribe

    $0.18 per hour

    pay-as-you-go

    • Streaming and non-streaming priced the same
  • Segment Anything Model 3.1

    $2.50 per 1,000 images

    pay-as-you-go

    • Video at $0.20 per 1,000 frames

Plans as listed on the official pricing page on . Prices can change and may vary by country or exclude tax. Check the official pricing page.

Meta Model API alternatives

The main Meta Model API alternatives listed on Debuly are OpenAI API, Gemini API, Mistral AI, Cohere, Together AI and DeepInfra. The table compares pricing and platforms; order is A–Z.

Meta Model API compared with its alternatives
ToolPricingStarting pricePlatforms
Meta Model APIPaid—API, Web
CohereFreemium—API, Web, Private deployment
DeepInfraPaid—API, Web
Gemini APIFreemium—Web, API
Mistral AIFreemium$14.99/moWeb, API, CLI, Self-hosted
OpenAI APIPaid$0.1, usage-basedAPI, Web
Together AIPaid—Web, API
  1. Cohere

    Enterprise AI company offering Command, Embed and Rerank models by API, the North agent platform and private deployments.

  2. DeepInfra

    AI inference cloud with an OpenAI-compatible API for 100+ open models, plus private GPU deployments, GPU rental and hosted agents.

  3. Gemini API

    Google's developer API for calling Gemini models from your own apps, with a free tier, pay-as-you-go pricing and Google AI Studio.

  4. Mistral AI

    AI company offering open-weight models, the Vibe assistant, Studio and an API, with a free plan and paid tiers.

  5. OpenAI API

    Developer API for OpenAI's GPT models, image generation, realtime voice and transcription, priced per million tokens with batch and flex options.

  6. Together AI

    AI cloud for running open-source models through serverless and batch inference, dedicated endpoints, GPU clusters and fine-tuning.

Meta Model API FAQ

How much does Meta Model API cost?

Muse Spark is $1.25 per 1M input tokens, $0.15 per 1M cached input tokens and $4.25 per 1M output tokens. Muse Wonder is $4.00, $0.20 and $20.00. Muse Image is $0.01 per image.

Is there a free tier?

The rate-limit table lists a Free tier at 60 requests and 2,000,000 tokens per minute, and a Paid tier at 3,000 requests and 4,000,000 tokens per minute. Check the docs for current free usage terms.

What is Contributor mode?

Contributor mode prices Muse Spark lower, at $0.10 input and $0.20 output per 1M tokens, and in return Meta gets permission to train future models on your content.

Can I use the OpenAI SDK with Meta Model API?

Yes. The site says you can generate an API key and point existing OpenAI SDK compatible code at it. Muse Spark is also available on OpenRouter.

Does Meta Model API support web search?

Yes. The web_search tool grounds answers in live results. It is billed at $2.50 per 1,000 searches on top of normal token charges.

What are alternatives to Meta Model API?

Comparable model APIs include the OpenAI API, Google Gemini API, Anthropic API, Mistral AI and Cohere.

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

SourceTypeAccessed
dev.meta.ai/official
dev.meta.ai/docs/pricing-rate-limitspricing
dev.meta.ai/docs/getting-started/pricing-rate-limitspricing
dev.meta.ai/products/meta-model-apiofficial
dev.meta.ai/docs/modelsdocs

Image credits

  • Logo: from dev.meta.ai · fetched · Vendor trademark, used only to identify the product; removed on request
  • Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.