Muse Spark 1.1
Muse Spark 1.1 is a chat-oriented large language model from Meta, listed in the ComputePrices catalog with availability and pricing tracked across inference providers.
API Pricing
| Provider | Input / 1M | Output / 1M | Cached / 1M |
|---|---|---|---|
| $1.25 | $4.25 | $1.00 | |
| $1.25 | $4.25 | $0.150 |
Prices updated daily. Last check: Oct 15, 2026
Muse Spark 1.1 pricing by provider
Input, output, and batch rates, plus alternatives, for one provider at a time.
Performance & Benchmarks
Source: Artificial Analysis →Reasoning & Knowledge
- GPQA Diamond89.8%
- Humanity's Last Exam46.2%
Coding
- SciCode58.8%
Agentic & Tool Use
- Terminal-Bench v2.177.9%
- τ-bench Banking31.8%
Instruction & Long Context
- Long-Context Reasoning77.7%
Benchmarks measured Oct 2026. Scores are independent evaluations, not vendor-reported.
Model Details
General
- Creator
- Meta
- Modalities
- Text
Capabilities
- Open Source
- No
Strengths & Limitations
Strengths
- Chat-tuned interface suitable for instruction following and multi-turn conversation without additional scaffolding
- Incremental 1.1 revision within the Muse Spark line, so prompts and integrations built for the line generally carry over
- Attributed to Meta, a creator with an established track record of publishing language models
- Served through inference API providers, so it can be compared side by side with alternatives in the pricing table on this page
- Text generation workloads can be routed to it without provisioning or managing GPU infrastructure
Limitations
- Context window length is not recorded in our database, so long-document workloads need verification with the provider
- No verified benchmark scores are available for this model in our data set
- Throughput and time-to-first-token are untracked — the Artificial Analysis figures we hold are placeholders of zero, not measurements
- Multimodal capabilities, tool calling, and structured output support are not confirmed in our metadata
- Provider availability may be narrower than for more widely distributed Meta models, limiting failover options
Key Features
About Muse Spark 1.1
Common Use Cases
Muse Spark 1.1 fits general text chat workloads: answering questions, drafting and rewriting copy, summarizing supplied text, and powering assistant-style interfaces where the application sends a conversation history and receives a text reply. Because we do not have confirmed context window or multimodal metadata for it, it is best matched to prompts of moderate length and text-only inputs unless your provider documents otherwise. For workloads that hinge on very long documents, image understanding, or verified reasoning benchmarks, run a direct evaluation against models in our catalog with confirmed specifications before committing. Teams evaluating cost-per-token for high-volume text generation can use the pricing table on this page to see which providers currently serve Muse Spark 1.1 and at what rates.
Frequently Asked Questions
How much does Muse Spark 1.1 cost to use?
Pricing varies by provider and by pricing type — input tokens, output tokens, and any cached or batch rates are billed separately and change over time. See the pricing table on this page for current per-provider rates rather than relying on a fixed figure.
What is Muse Spark 1.1 best used for?
It is a chat model, so it suits conversational assistants, question answering, drafting and editing text, and summarization of supplied content. Given that we do not have confirmed long-context or multimodal specifications for it, text-focused workloads of moderate prompt length are the safest fit.
Who created Muse Spark 1.1?
Our catalog attributes Muse Spark 1.1 to Meta. The 1.1 version number indicates a point revision within the Muse Spark line rather than a separate model family.
What is the context window of Muse Spark 1.1?
We do not currently have a confirmed context window recorded for this model. Check the documentation of the specific inference provider you plan to use, since maximum input and output lengths can also differ between hosts serving the same model.
Why are the speed benchmarks showing zero?
The output tokens per second and time-to-first-token values we hold from Artificial Analysis are currently reported as zero, which means the model has not been measured on that harness. It should not be read as a performance result. Provider-reported latency or your own load testing is the better guide until measured figures are available.