Muse Spark Open
Model ranking
Not matched to the current verified ranking snapshot.
| Provider | Price | Provider model ID | Status | Context / max output | Source |
|---|---|---|---|---|---|
Router catalog Tariff not listedInput $1.25USD / 1M tokensOutput $4.25USD / 1M tokens Updated 2026-11-22 · 15:43 UTCOpenRouter · Router catalogInput: $1.25 Output: $4.25 Cache Read: $0.15 OpenRouter route catalog price, not a direct-provider price. Routing, modality and endpoint conditions can change the actual bill. Checked 2026-11-22T15:43:15.988Z Price source ↗ | meta/muse-spark-open | Catalog-listedConditionsOpenRouter catalog quote; upstream conditions and regional availability can differ. Open weights use the Muse Model License and its usage conditions. They should not be described as unrestricted Apache or MIT weights. Large parameter counts also mean a downloaded checkpoint needs substantial infrastructure. | Context tokens 1048576Output tokens 131072 | Official source ↗2026-10-31 |
Benchmark review · November 19, 2026. The latest verified Artificial Analysis Intelligence Index score is 25 for Muse Spark Open Xhigh, checked on November 19, 2026. Rank is among verified models in this curated catalog. Benchmark source.
Muse Spark Open is Meta’s open-weight model for coding and agentic workflows. The released architecture has 816 billion total parameters and 42 billion active parameters, with native multimodal input and text output.
Muse Spark Open specifications and access
Open weights use the Muse Model License and its usage conditions. They should not be described as unrestricted Apache or MIT weights. Large parameter counts also mean a downloaded checkpoint needs substantial infrastructure.
The specification table distinguishes manufacturer references from provider deployment limits. Check the linked endpoint before choosing a maximum prompt or response size. If a field is unpublished, treat it as unknown rather than assuming the limit of another model in the family.
Muse Spark Open pricing and cost planning
The verified OpenRouter route lists $1.25 input, $4.25 output and $0.15 cached input per million tokens. This is a hosted route quote. Downloading weights does not make GPU operation, storage or serving free.
For a practical budget, count input, generated output and repeated attempts separately. Also record tool charges and caching conditions. Compare several complete tasks with the same acceptance criteria. A low input price alone can hide expensive long answers or additional retries.
Where to use Muse Spark Open
Compare hosted inference with self-hosting when deployment control matters. Record the quantization, hardware allocation and concurrency in every test, because a model name alone does not describe serving cost or speed.
Start with a small evaluation set containing easy, difficult and incomplete requests. Keep the expected result for each task. Then inspect accuracy, response time and the effort needed to correct failures. Use those observations to decide whether a more capable or less expensive route fits the workload.
Muse Spark Open benchmark and alternatives
The current comparison uses the Artificial Analysis Intelligence Index v5.0 snapshot checked on October 31, 2026. The recorded configuration is xhigh. Its score describes that evaluation setting; a different reasoning budget or tool setup can change the result.
Review the official Muse Spark Open reference and the technical documentation. Use the benchmark evaluation to check the tested configuration.
Compare Muse Spark Open with other entries in the AI model rankings and pricing directory. For an integration, inspect the linked provider offers and run your own acceptance tests before choosing a production endpoint.
Specifications
- Catalog model ID
- meta/muse-spark-open
- Developer model ID / revision
- Muse Spark Open
- Input modalities
- text, image, video, file, audio
- Output modalities
- text
- Context / input token limit
- 1,000,000
- Maximum output tokens
- 131,072
- Tool calling
- Supported
- Structured output
- Supported
- API parameters (source-specific)
- frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p
- Weight access
- Open weights
- Weight license / access terms
- Muse Model License
- Total parameters (billions)
- 816
- Active parameters (billions)
- 42
- Weight format / quantization
- BF16 and NVFP4 official checkpoints; deployment-dependent
- Download size (GB)
- Depends on checkpoint and precision; no verified file size
- Release date
- 2026-10-15
- Notes
- Official architecture counts are 816B total and 42B active. Context is published as 1M by Meta; route-specific exact limits are shown on each offering. Download size varies by precision and is not inferred from parameter count. OpenRouter context, maximum output and supported parameters describe that route; they are not universal limits across every provider.
Catalog entries link to their specification and availability sources. This is a curated selection, not every model each provider offers. Best prices are the lowest collected USD tariffs, not a total-cost estimate. Token offers are selected by input price, with output from the same tariff; workload mix may change the cheapest provider. Listed variants may require batch processing, off-peak hours, or specific media settings. Price coverage varies by provider; collected tariffs include their source and observation time. Arena user-preference scores are checked automatically every six hours against the latest published Text / overall dataset. Source: Arena / LMArena; CC BY 4.0. Selection and catalog matching by XPU LIVE. They apply to the shown evaluation configuration; listed prices may use different settings. Models without verified scores are not assumed to be weaker.