Muse Spark Open
Muse Spark Open is a large language model made by Meta, added to the catalog on 15 Oct 2026. It has a context window of 1,048,576 tokens and costs $1.25 per 1M input tokens and $4.25 per 1M output tokens. It accepts text, images, files, audio and video and produces text. Typical uses: agentic coding workflows, coding tasks involving tool use and working with lengthy coding task context.
Price checked ยท source: OpenRouter
What Muse Spark Open is
Meta: Muse Spark Open is an open-weight model optimized for coding and described as trained specifically for agentic coding workflows. It accepts text, images, video, files, and audio, and produces text. The catalog lists support for tools, structured outputs, response formats, and reasoning. These capabilities make its stated scope relevant to coding work that uses tools, requests a defined response format, or draws on more than one kind of input.
It has a context length of 1,048,576 tokens and a maximum output of 943,718 tokens. The context limit gives the model room to receive extensive task material in one interaction, while the output limit allows lengthy text responses. Those limits may be relevant when coding tasks involve large amounts of code or supporting documentation. The provider description identifies deep codebase understanding as a focus, but does not name programming languages, repositories, or development environments.
Muse Spark Open returns text rather than generated images, audio, or video. Although it can accept those media types as inputs, the listed output modality is text. Its combination of coding focus, reasoning support, and tool capability may suit development workflows where an assistant needs to examine provided context and respond in text. Structured outputs and response-format support can also fit tasks where the requested response needs a defined structure. The catalog does not specify particular tools, schemas, or coding integrations, so those details depend on the intended setup.
Key features
- Accepts text, image, video, file, and audio inputs
- Produces text output
- Has a context length of 1,048,576 tokens
- Supports tools, structured outputs, response formats, and reasoning
- Described as optimized for agentic coding workflows
Typical uses
- Agentic coding workflows
- Coding tasks involving tool use
- Working with lengthy coding task context
- Generating structured text responses
Muse Spark Open pricing
US dollars, list price through OpenRouter.
| Item | Standard |
|---|---|
| Input, per 1M tokens | $1.25 |
| Output, per 1M tokens | $4.25 |
| Cached input read, per 1M tokens | $0.15 |
No price change since it was added on 15 Oct 2026.
| What a task costs | Cost |
|---|---|
| Answer a short question500 in / 300 out | $0.0019 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.035 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.046 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.194 |
Muse Spark Open specifications
| Model id | meta/muse-spark-open |
|---|---|
| Version | muse-spark-open-20261015 |
| Context window | 1,048,576 tokensAbout 786,432 English words |
| Max output | 943,718 tokens |
| Knowledge cutoff | Not published |
| Added to catalog | 15 Oct 2026 |
| Input types | text, image, video, file, audio |
|---|---|
| Output types | text |
| Tool calling | Yes |
| Structured outputs | Yes |
| Reasoning effort control | Yes |
| Content moderation | Not applied by provider |
Sample outputs
The same prompts are given to every model, and the unedited output is shown with the date it was produced. Upend does not publish leaderboard scores it did not run.
No output has been recorded for Muse Spark Open yet.
Compare Muse Spark Open
Side by side with other models: prices, context window, what each accepts and supports, and the same four tasks priced on both.