GPT-6 Cyber
OpenAI's restricted cybersecurity model helps approved researchers discover vulnerabilities, validate exploits and develop patches within authorized scope.
Muse Ember
dev.meta.ai
Meta’s multimodal model handles repeated coding subtasks, tool calls, routing and document processing inside agent workflows.
Visit website
Pricing
Paid
Available on
Meta Model API, Muse Code
Category
Llm Models
What is Muse Ember?
Muse Ember is Meta’s model for high-volume execution within agent workflows. It serves developers processing recurring tasks such as coding subtasks, request routing, extraction and long-document parsing.
Teams can combine its multimodal processing with larger Muse models that handle planning and orchestration. Its focus is repeated, bounded work across coding and information-processing pipelines.
Key features
Multimodal input
Accepts text, images, video, audio and PDFs, with text responses.
Long-context processing
Provides a 1,048,576-token context window and a maximum output of 524,288 tokens.
Agent integration
Supports tool calling, computer use, image understanding and search grounding through configured integrations.
Structured responses
Supports schema-constrained JSON for routing and extraction, with Responses, Chat Completions and Messages endpoints available.
Who is it for?
Support operations
Classify incoming tickets and extract fields for downstream handling.
Document processing
Summarize documents and organize extracted information into structured records.
Agent development
Assign recurring coding subtasks and tool calls to a dedicated execution model.
How to use Muse Ember
- Create a Meta Model API key through the developer setup.
- Configure a compatible client with the Model API endpoint and your key.
- Select muse-ember-1.0 and send a representative task with its inputs.
- Configure tools or structured output as needed, then review responses before integrating the workflow.
Muse Ember pricing
Paid
Meta Model API uses pay-as-you-go token billing with no minimum or upfront commitment. Muse Ember uses Standard mode, with separate long-context rates above 200K input tokens. Consult the official pricing page for current rates.
Check current pricing ↗What to know before you use it
- Responses contain text, including when inputs contain audio, video or images.
- Meta recommends Muse Spark or Muse Wonder for difficult single-shot coding, knowledge-heavy questions and visual coding.
- Requests exceeding 200K input tokens incur long-context rates for the entire request. Cached input and stored conversation history count toward this threshold.
- Published API limits are 3,000 requests per minute and 4,000,000 tokens per minute. Search grounding carries a separate usage charge.
Frequently asked questions
Can I use Muse Ember in an existing API integration?
It supports Responses, Chat Completions and Messages. Configure a compatible client through Meta’s quickstart and select muse-ember-1.0.
Can I access it through Muse Code?
Muse Ember is available through pay-as-you-go access and every Muse Code subscription plan. Requests above 200K input tokens consume subscription usage faster.
How does a long conversation affect API billing?
Stored conversation history contributes to input length. When total input exceeds 200K tokens, higher long-context rates apply to the whole request.