Muse Spark Open
Summary
Muse Spark Open is an open-weight language model from Meta, released on 15 Oct 2026. It accepts text, images, video, PDFs and audio and generates text, with a 1M-token context window and up to 944K output tokens per response. As of 17 Oct 2026, the lowest listed API price is $1.25 per million input tokens and $4.25 per million output tokens (Fireworks AI), across 4 providers we track.
Open Llama multimodal model for image understanding and text reasoning
Last verified 17 Oct 2026. Every figure on this page links to its source.
Specifications
| Developer | Meta |
|---|---|
| Release date | 15 Oct 2026 [models.dev] |
| Status | Generally available [models.dev] |
| Weights | Open weights [Hugging Face] |
| License | muse-model-license [Hugging Face] |
| Parameters | 822B [Hugging Face] |
| Context window | 1M (1,048,576 tokens) [models.dev] |
| Max output | 944K tokens [models.dev] |
| Knowledge cutoff | 4 Jan 2026 [models.dev] |
| Input | Text, images, video, PDFs and audio [models.dev] |
| Output | Text [models.dev] |
| Reasoning mode | Yes [models.dev] |
| Tool calling | Yes [models.dev] |
| Structured output | Yes [models.dev] |
| Hugging Face | meta-models/Muse-Spark-Open-816B-A42B [Hugging Face] |
Muse Spark Open API pricing by provider
USD per million tokens. Sorted by input price.
| Provider | Input | Output | Cached input | Context | Source |
|---|---|---|---|---|---|
| Fireworks AI accounts/fireworks/models/muse-spark-open | $1.25 | $4.25 | $0.15 | 1M | models.dev, 17 Oct 2026 |
| Nebius AI Studio meta-models/muse-spark-open | $1.25 | $4.25 | $1.25 | 1M | models.dev, 17 Oct 2026 |
| OpenRouter meta/muse-spark-open | $1.25 | $4.25 | $0.15 | 1M | OpenRouter, 17 Oct 2026 |
| Vercel AI Gateway meta/muse-spark-open | $1.25 | $4.25 | $0.15 | 1M | models.dev, 17 Oct 2026 |
Running Muse Spark Open locally
Estimated memory, from the parameter count on Hugging Face. Includes about 0.8 GB of runtime overhead.
| Quantization | Weights | Total, 8K context | Total, 32K context |
|---|---|---|---|
| Q4_K_M (4-bit) | 498 GB | 499 GB | – |
| Q8_0 (8-bit) | 873 GB | 874 GB | – |
| FP16 / BF16 (16-bit) | 1644 GB | 1645 GB | – |
At 4-bit with an 8K conversation, Muse Spark Open needs about 499 GB, more than any single consumer graphics card; it needs a unified-memory machine, a multi-GPU setup or a split between graphics card and system RAM. The model's architecture is not public, so the conversation cache is not included. Check your own hardware.
Frequently asked questions
What is the context window of Muse Spark Open?
Muse Spark Open has a context window of 1,048,576 tokens (1M) and can generate up to 943,718 tokens in a single response, according to models.dev.
How much does the Muse Spark Open API cost?
Muse Spark Open is listed from $1.25 per million input tokens and $4.25 per million output tokens (Fireworks AI). Prices last verified 17 Oct 2026.
Is Muse Spark Open open source?
Muse Spark Open is an open-weight model: its weights are published on Hugging Face as meta-models/Muse-Spark-Open-816B-A42B under the muse-model-license license. Check the license terms before commercial use.
When was Muse Spark Open released?
Meta released Muse Spark Open on 15 Oct 2026, according to models.dev.
Which providers offer Muse Spark Open?
We track Muse Spark Open on 4 providers: Fireworks AI, Nebius AI Studio, OpenRouter, Vercel AI Gateway.
How much VRAM does Muse Spark Open need?
About 499 GB at 4-bit (Q4_K_M) with an 8K-token conversation, and about 874 GB at 8-bit. That is more than any single consumer graphics card holds. These are estimates from the model's published parameter count.
Change history
- License changed from none to muse-model-license
- Parameter count changed from none to 822B
- Now listed on OpenRouter
- Added to the directory
- Now listed on Fireworks AI
- Now listed on Nebius AI Studio
- Now listed on Vercel AI Gateway
Sources
- https://openrouter.ai/meta/muse-spark-open
- https://github.com/sst/models.dev/blob/dev/providers/openrouter/models/meta/muse-spark-open.toml
- https://huggingface.co/meta-models/Muse-Spark-Open-816B-A42B
- https://github.com/sst/models.dev/blob/dev/providers/fireworks-ai/models/accounts/fireworks/models/muse-spark-open.toml
- https://github.com/sst/models.dev/blob/dev/providers/nebius/models/meta-models/muse-spark-open.toml
- https://github.com/sst/models.dev/blob/dev/providers/vercel/models/meta/muse-spark-open.toml