Muse Wonder
Muse Wonder is a chat model from Meta tracked in our catalog with third-party throughput measurements from Artificial Analysis.
API Pricing
| Provider | Input / 1M | Output / 1M | Cached / 1M |
|---|---|---|---|
| $4.00 | $20.00 | $0.200 |
Prices updated daily. Last check: Oct 13, 2026
Muse Wonder pricing by provider
Input, output, and batch rates, plus alternatives, for one provider at a time.
Performance & Benchmarks
Source: Artificial Analysis →Reasoning & Knowledge
- Humanity's Last Exam50.1%
Coding
- SciCode60.3%
Instruction & Long Context
- Long-Context Reasoning80.7%
Benchmarks measured Oct 2026. Scores are independent evaluations, not vendor-reported.
Model Details
General
- Creator
- Meta
- Modalities
- Text
Capabilities
Strengths & Limitations
Strengths
- Measured output throughput of approximately 81.4 tokens per second, sufficient for streamed conversational output
- Independent serving benchmarks available from Artificial Analysis rather than creator-published figures alone
- Attributed to Meta, a creator with an established history of releasing and serving language models
- Steady generation rate means long-form completions finish in predictable time once streaming starts
- Chat-formatted model, so it slots into standard conversational request/response integrations
- Latency profile is well documented, which makes it straightforward to decide whether it fits a given workload
Limitations
- Measured time to first token of roughly 28.6 seconds makes it unsuitable for latency-sensitive interactive use
- No confirmed context window in our data, so long-document suitability cannot be assessed
- We do not track confirmed modality support, tool calling, or structured output for this model
- No published benchmark accuracy scores in our dataset to compare reasoning or coding quality against peers
- Provider availability is limited, which narrows pricing and routing options compared with widely hosted models
Key Features
About Muse Wonder
Common Use Cases
Given a measured first-token latency near 29 seconds paired with a reasonable ~81 tokens/second generation rate, Muse Wonder fits asynchronous and batch-style workloads far better than real-time chat. Practical candidates include background document drafting, offline content generation, scheduled summarization jobs, queued agent steps where a human is not waiting on the response, and evaluation or research pipelines where throughput per request matters more than responsiveness. Conversely, customer-facing chatbots, IDE autocomplete, voice interfaces, and any interactive surface with a sub-second expectation would be poorly served by this latency profile. Because we do not have confirmed context window or modality data, validate those requirements with your provider before committing the model to a pipeline that depends on long inputs or non-text data.
Frequently Asked Questions
How much does Muse Wonder cost to run?
Pricing depends on which provider is serving the model and whether you are billed per input token, per output token, per request, or through a dedicated capacity arrangement. Rates also change frequently as providers adjust. Check the pricing table on this page for current per-provider figures.
What is Muse Wonder best used for?
Asynchronous and batch text generation — background drafting, queued summarization, offline content pipelines, and research or evaluation runs. Its measured ~28.6 second time to first token makes it a poor fit for interactive chat, autocomplete, or voice applications.
Why is the time to first token so high?
A ~28.6 second first-token latency, as measured by Artificial Analysis, is well above typical chat endpoints. That pattern usually reflects extended processing before the first visible token, request queueing, or constrained serving capacity. The measurement reflects conditions at benchmark time and may differ on your provider and region.
Is Muse Wonder fast once it starts responding?
Yes, relative to its startup cost. The measured output rate of about 81.4 tokens per second is a normal interactive streaming speed, so a long completion progresses at a reasonable pace — the delay is concentrated in getting the first token out.
What context window does Muse Wonder support?
We do not have a confirmed context window for Muse Wonder in our database. A missing value here means unverified, not unlimited or small — confirm the supported input length directly with the provider you plan to use.