Meta model product
Muse Spark is a natively multimodal reasoning model developed by Meta Superintelligence Labs.
Updated Sep 8, 2026. Default version: Muse Spark
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Benchmark scores for Muse Spark.
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkFrontierScience Research | Score38.30% | Rank01 | Participants4 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkHealthBench Hard | Score42.80% | Rank01 | Participants9 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkIPhO 2025 | Score82.60% | Rank01 | Participants2 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkMedXpertQA | Score78.40% | Rank01 | Participants12 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkZEROBench | Score0.33 points | Rank03 | Participants10 | Percentile77.78% | EvidenceC | Evaluated |
| BenchmarkLiveCodeBench Pro | Score0.80 points | Rank04 | Participants4 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkSimpleVQA | Score0.71 points | Rank04 | Participants14 | Percentile76.92% | EvidenceC | Evaluated |
| BenchmarkScreenSpot Pro | Score84.10% | Rank05 | Participants26 | Percentile84.00% | EvidenceC | Evaluated |
| BenchmarkLM Arena Vision Style Control | Score1,293.73 rating | Rank05 | Participants107 | Percentile96.23% | EvidenceA | Evaluated |
| BenchmarkLM Arena Vision | Score1,306.23 rating | Rank07 | Participants107 | Percentile94.34% | EvidenceA | Evaluated |
| BenchmarkHumanity's Last Exam | Score58.40% | Rank08 | Participants103 | Percentile93.14% | EvidenceC | Evaluated |
| BenchmarkDeepSearchQA | Score74.80% | Rank09 | Participants10 | Percentile11.11% | EvidenceC | Evaluated |
| BenchmarkARC-AGI v2 | Score42.50% | Rank10 | Participants18 | Percentile47.06% | EvidenceC | Evaluated |
| BenchmarkERQA | Score64.70% | Rank11 | Participants26 | Percentile60.00% | EvidenceC | Evaluated |
| BenchmarkLM Arena Text Style Control | Score1,488.17 rating | Rank11 | Participants210 | Percentile95.22% | EvidenceA | Evaluated |
| BenchmarkMMMU-Pro | Score80.40% | Rank13 | Participants69 | Percentile82.35% | EvidenceC | Evaluated |
| BenchmarkCharXiv-R | Score86.40% | Rank14 | Participants55 | Percentile75.93% | EvidenceC | Evaluated |
| BenchmarkLM Arena Document Style Control | Score1,466.33 rating | Rank14 | Participants33 | Percentile59.38% | EvidenceA | Evaluated |
| BenchmarkTau2 Telecom | Score91.50% | Rank16 | Participants36 | Percentile57.14% | EvidenceC | Evaluated |
| BenchmarkLM Arena Text | Score1,473.42 rating | Rank18 | Participants210 | Percentile91.87% | EvidenceA | Evaluated |
| BenchmarkLM Arena Document | Score1,443.25 rating | Rank20 | Participants33 | Percentile40.63% | EvidenceA | Evaluated |
| BenchmarkTerminal-Bench 2.0 | Score59.00% | Rank24 | Participants51 | Percentile54.00% | EvidenceC | Evaluated |
| BenchmarkLM Arena Text Factuality | Score1,466.50 rating | Rank24 | Participants121 | Percentile80.83% | EvidenceA | Evaluated |
| BenchmarkSWE-Bench Verified | Score77.40% | Rank27 | Participants113 | Percentile76.79% | EvidenceC | Evaluated |
| BenchmarkGPQA | Score89.50% | Rank31 | Participants247 | Percentile87.80% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Pro | Score52.40% | Rank46 | Participants55 | Percentile16.67% | EvidenceC | Evaluated |
Preference and agent-evaluation results for the default version.
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenatext style control | Categoryfrench | Rank02 | Rating / score1,524.77 | Votes468 | ObservationsN/A | Result date |
| Arenavision | Categoryhumor | Rank02 | Rating / score1,346.00 | Votes227 | ObservationsN/A | Result date |
| Arenavision style control | Categoryhumor | Rank02 | Rating / score1,325.94 | Votes227 | ObservationsN/A | Result date |
| Arenatext style control | Categorygerman | Rank03 | Rating / score1,508.70 | Votes235 | ObservationsN/A | Result date |
| Arenatext style control | Categorykorean | Rank04 | Rating / score1,473.97 | Votes248 | ObservationsN/A | Result date |
| Arenavision style control | Categoryenglish | Rank05 | Rating / score1,288.53 | Votes2,329 | ObservationsN/A | Result date |
| Arenavision style control | Categoryoverall | Rank05 | Rating / score1,293.73 | Votes5,585 | ObservationsN/A | Result date |
| Arenatext | Categoryfrench | Rank06 | Rating / score1,501.62 | Votes468 | ObservationsN/A | Result date |
| Arenatext | Categorygerman | Rank06 | Rating / score1,495.01 | Votes235 | ObservationsN/A | Result date |
| Arenavision | Categorycreative writing vision | Rank06 | Rating / score1,327.48 | Votes334 | ObservationsN/A | Result date |
| Arenavision style control | Categorycreative writing vision | Rank06 | Rating / score1,303.49 | Votes334 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry legal and government | Rank07 | Rating / score1,503.74 | Votes974 | ObservationsN/A | Result date |
| Arenavision | Categoryenglish | Rank07 | Rating / score1,302.94 | Votes2,329 | ObservationsN/A | Result date |
| Arenavision | Categoryoverall | Rank07 | Rating / score1,306.23 | Votes5,585 | ObservationsN/A | Result date |
| Arenavision style control | Categoryocr | Rank08 | Rating / score1,301.78 | Votes4,021 | ObservationsN/A | Result date |
| Arenatext | Categorykorean | Rank09 | Rating / score1,459.84 | Votes248 | ObservationsN/A | Result date |
| Arenatext style control | Categoryenglish | Rank09 | Rating / score1,495.50 | Votes6,471 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry business and management and financial operations | Rank09 | Rating / score1,490.07 | Votes2,747 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry medicine and healthcare | Rank10 | Rating / score1,503.00 | Votes955 | ObservationsN/A | Result date |
| Arenatext style control | Categoryspanish | Rank10 | Rating / score1,481.55 | Votes501 | ObservationsN/A | Result date |
| Arenavision style control | Categorychinese | Rank10 | Rating / score1,342.05 | Votes450 | ObservationsN/A | Result date |
| Arenatext | Categoryindustry legal and government | Rank11 | Rating / score1,489.12 | Votes974 | ObservationsN/A | Result date |
| Arenatext style control | Categoryoverall | Rank11 | Rating / score1,488.17 | Votes13,572 | ObservationsN/A | Result date |
| Arenavision | Categoryocr | Rank11 | Rating / score1,306.97 | Votes4,021 | ObservationsN/A | Result date |
| Arenatext factuality | Categorycreative writing | Rank12 | Rating / score1,463.87 | Votes1,571 | ObservationsN/A | Result date |
| Arenatext style control | Categoryexclude ties | Rank12 | Rating / score1,500.93 | Votes10,291 | ObservationsN/A | Result date |
| Arenatext style control | Categoryhard prompts english | Rank12 | Rating / score1,507.61 | Votes4,353 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry entertainment and sports and media | Rank12 | Rating / score1,462.19 | Votes2,510 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry software and it services | Rank12 | Rating / score1,519.97 | Votes5,318 | ObservationsN/A | Result date |
| Arenatext style control | Categorynon english | Rank12 | Rating / score1,476.97 | Votes7,101 | ObservationsN/A | Result date |
Official vendor API pricing appears first, followed by individual provider offers.
The default version has no current input or output token prices.
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Provider-specific output speed and catalog latency for Muse Spark. Runtime does not affect the capability score.
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Technical details for the model's default version.
Available versions of this model. The score column identifies the version used in the overall ranking.
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionMuse Spark | Released | LLMBoard72.09 | ParametersN/A | ContextN/A | Max outputN/A | Open weightsNo | LicenseProprietary |
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
Key information about Muse Spark and its available data.
Muse Spark is the first model in the Muse family, with support for tool-use, visual chain of thought, and multi-agent orchestration. Its Contemplating mode orchestrates multiple agents reasoning in parallel and achieved 58% on Humanity's Last Exam and 38% on FrontierScience Research.
Data as of 2026-09-08.
Common questions about Muse Spark.
Muse Spark's default version was released on Apr 8, 2026.
No official standard PAYG price is currently available for Muse Spark.
Muse Spark was created by Meta.
A context window is not available for the default version.
No. The default version is not marked as having publicly available weights.
No provider offering is currently linked to the default version.