Muse Spark 1.3

Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2. It has a context window of 1,048,576 tokens and a maximum output of 131,072 tokens. It is not open weights. It supports reasoning mode and tool calling. It accepts audio, image, pdf, text, and video, and returns text. Pricing starts at $1.25 per million input tokens and $4.25 per million output tokens, with cached input at $0.15 per million tokens.

Same headline figures as Muse Spark 1.2.

Range across the hosts that serve it: up to $1.38 per 1M input tokens.

• ·Long-horizon agent collaboration for multi-agent workflows
• ·Improved instruction following relative to Muse Spark 1.2
• ·Coding efficiency for long-running workflows
• ·Accepts audio, image, pdf, text, and video

• Reach for it for long-running multi-agent collaboration
• Reach for it for coding workflows with large context
• Reach for it for multimodal reasoning across text, image, audio, video, and pdf

10 hosts serve this model at the price above · the maker's documentation

The registry keeps the six most recent versions of a line. Older ones stay in the file and can be surfaced without a rebuild.

What is the context window?

The context window is 1,048,576 tokens.

What is the pricing?

Input from $1.25 per million tokens, output from $4.25 per million, cached input from $0.15 per million.

Is the model open weights?

No, the weights are not open.