Sponsored by Byond Boundrys Consulting - Empowering Ideas, Delivering Results
Meta Advanced

Muse Spark 1.3

Muse Spark 1.3 is Meta’s model for long-horizon agentic coding, computer use and multi-agent workflows, available in public preview through the Meta Model API.

Multimodal ModelsReasoning ModelsTextImageVideo Paid
In plain English

What is this model and why does it matter?

Muse Spark 1.3 is Meta's newest Muse Spark model for long-horizon agentic workflows, coding, multimodal perception, computer use and coordinated multi-agent execution.

Agentic codingLong-horizon tasksComputer useMultimodal reasoningMulti-agent workflows
Model overview

Muse Spark 1.3: features, use cases and important details

Muse Spark 1.3 is a current AI model verified from first-party Meta sources.

Muse Spark 1.3 verified specifications

Muse Spark 1.3 is Meta’s newest Muse Spark model for long-horizon agentic workflows, coding, multimodal perception, computer use and coordinated multi-agent execution. Its verified context or usage limit is Long-context agentic workflow; exact 1.3 token limit is not stated on Meta’s public product page, with Not publicly stated on the verified first-party page maximum output.

Muse Spark 1.3 pricing and access

Muse Spark 1.3 is available through Meta Model API public preview. Meta describes the API as self-serve and affordable, but the verified public product page does not publish a fixed token rate.

Muse Spark 1.3 best uses

Agentic coding, Long-horizon tasks, Computer use, Multimodal reasoning, Multi-agent workflows.

Muse Spark 1.3 limitations

Preview behavior may change, Production limits should be confirmed in Meta Model API, Complex agents still need supervision.

Get started

How to use this model

  1. Open Meta Model API access.
  2. Select Muse Spark 1.3.
  3. Provide text, image, video or document context.
  4. Connect tools or subagents when required.
  5. Evaluate latency, tool behavior and cost before production.
Copy and try

Example prompts

  • Review this repository and coordinate a multi-agent implementation plan.
  • Analyze this video and build the required interface.
  • Use computer tools to complete this long-running workflow.
Capabilities

What it can do

  • Agentic coding
  • Multimodal perception
  • Computer use
  • Tool use
  • Multi-agent orchestration
  • Long-horizon workflows
Best for

Practical use cases

  • Coding agents
  • Research agents
  • Computer-use automation
  • Multimodal analysis
  • Software development
Pricing

What does it cost?

Muse Spark 1.3 is available through Meta Model API public preview. Meta describes the API as self-serve and affordable, but the verified public product page does not publish a fixed token rate.

InputCheck current Meta Model API billing
OutputCheck current Meta Model API billing
Simple summaryAPI access is usage-based; exact current token rates should be checked in Meta Model API billing.

What stands out

  • Newest Muse Spark generation
  • Strong coding focus
  • Multimodal
  • Built for long-running agents

Things to consider

  • Public preview
  • Exact public token pricing is not stated
  • Proprietary
Limitations

Important restrictions and trade-offs

  • Preview behavior may change
  • Production limits should be confirmed in Meta Model API
  • Complex agents still need supervision
SimplifyAITools verdict

Our editorial take

A high-interest September 2026 Meta model for developers following frontier coding and agentic model competition.

References

Primary sources

  1. Open source 1 ↗
  2. Open source 3 ↗
Keep exploring

Related AI models