Skip to main content
NewSeedance 2.5 is live→
ModelRunner
Model creator on ModelRunner
Meta logo

Meta

meta

Meta's 3 models on ModelRunner span Audio-to-Audio, Image-to-Image and Sound, from $0.001667 per second of audio upward.

3 public models — every one runnable in a free playground, through the REST API and JavaScript SDK, or from an AI assistant over MCP.

Meta models & pricing

SAM Audio — Separate

SAM Audio — Separate

meta

Isolate any sound from an audio mixture by describing it in plain language.

$0.001667 per second of audioaudioaudiosegmentation
Segment Anything 2 (Auto-Segment)

Segment Anything 2 (Auto-Segment)

meta

Automatically segment a photo into a combined object/region mask — no prompts, points, or clicks needed.

$0.01 per imageimageimagesegmentationmask
musicgen

musicgen

meta

A fast, controllable auto-regressive Transformer for high-fidelity music generation.

≈$0.07 per audio cliptextaudio

What you can build with Meta models

Browse by capability

Call a Meta model in one request

Every endpoint shares the same lifecycle: submit to the queue, poll the status URL, fetch the result. Swap the endpoint to switch models — nothing else changes.

cURL

curl -X POST https://queue.modelrunner.run/meta/sam-audio/separate \
  -H "Authorization: Key $MRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "prompt": "..." }'
# → { "request_id": "...", "status_url": "...", "response_url": "..." }

Use Meta models from Claude & Cursor

Connect the ModelRunner MCP server once and your AI assistant can run every Meta model on this page as a tool — authorized via OAuth, results returned as hosted URLs in the conversation.

Claude Code

claude mcp add --transport http modelrunner https://mcp.modelrunner.run/mcp

Research, weights & licences

Frequently asked questions

What Meta models can I run on ModelRunner?

The catalog has 3 public Meta models: SAM Audio — Separate, Segment Anything 2 (Auto-Segment) and musicgen. They cover audio-to-audio, image-to-image and sound. Each one has its own page with a free playground, the full input schema, and copy-paste code.

How much do Meta models cost?

Pay-per-use, with no subscription: the one Meta model billed per second of audio costs $0.001667, and every page publishes its own rate — some are tiered, so a lower resolution or a shorter clip costs less per run. You buy credits and pay only for what you generate.

How do I call a Meta model from the API?

POST to https://queue.modelrunner.run/meta/sam-audio/separate with an Authorization: Key header and the input fields as the top-level JSON body (no wrapper object), then poll the returned status_url and fetch the result from response_url. The JavaScript SDK (@modelrunner/client) wraps that in a single subscribe() call, which is the one place an { input } object is used.

Can I run Meta models from Claude or Cursor?

Yes. Connect the ModelRunner MCP server (https://mcp.modelrunner.run/mcp) once and all 3 Meta models become tools your assistant can call with run_model — authorized over OAuth 2.1, so no API key goes into a config file. Results come back as hosted URLs in the conversation.

Where is the source code or paper for Meta models?

1 of the 3 Meta models in the catalog link their own upstream sources — musicgen — covering some combination of code repository, paper, weights and licence. The links are listed on this page and on each model's own page.

More on ModelRunner

Browse the full model catalog, compare live per-configuration rates on the pricing page, see curated sets on explore, or view every model creator.

Other creators in audio-to-audio