AIEvals

Models

Muse Spark 1.2

Specifications and API prices for Muse Spark 1.2, recorded by Artificial Analysis.

MetaReleased 5 August 2026Context 1,048,576 tokensOutput Not published in the checked sourcePrivate weights$1.25 in / $4.25 out per Mtok

Family Muse Spark39 benchmarks with a published result1 independent first place among the 46 models this site listsRead 17 October 2026

Where it stands

Rank compares tracked models with a score. Published shows the source’s rank and field size, when available.

BenchmarkScoreRankPublishedLeaderBehind
Vals IndexVals AI49.29%27of 3328of 41Gemini 4 Argon19.61 pts
Artificial Analysis Intelligence IndexArtificial Analysis4025of 42no list readClaude Opus 5.518 pts
Terminal-Bench 4.0Artificial Analysis7.07%36of 40no list readClaude Sonnet 5.556.57 pts
SWE-bench (Verified)Vals AI86.60%12of 2313of 88Claude Opus 510.40 pts
MMLU ProVals AI88.28%13of 2420of 138Claude Fable 5.14.10 pts
MMMU ProVals AI86.13%13of 2022of 93Claude Fable 5.14.51 pts
Vibe Code Bench v1.1Vals AI79.10%20of 3821of 106Claude Sonnet 5.513.29 pts
τ³-BankingArtificial Analysis34.85%22of 30no list readQwen3.8 Max16.49 pts
MLCR-AAArtificial Analysis31.11%15of 37no list readClaude Sonnet 5.543.89 pts
EnterpriseOps-Gym-AAArtificial Analysis47.27%8of 22no list readClaude Fable 53.85 pts
AutomationBench-AAArtificial Analysis40.61%34of 40no list readGemini 4 Argon36.90 pts
SciCodeArtificial Analysis57.41%13of 40no list readClaude Opus 5.59.49 pts
GDP.pdfArtificial Analysis17.40%27of 40no list readGPT-6.1 Sol13.60 pts
CritPtArtificial Analysis17.71%28of 42no list readGPT-5.6 Sol14.58 pts
AA-OmniscienceArtificial Analysis27.215of 42no list readClaude Opus 5.519.22
AA-LCR v1.1Artificial Analysis79.00%35of 42no list readKimi K39.67 pts
Terminal-Bench 2.1Vals AI69.66%25of 3528of 75Claude Opus 5.517.98 pts
ProofBench v1.1Vals AI43.00%27of 3431of 46Claude Sonnet 5.557.00 pts
LegalBenchVals AI85.26%14of 2721of 149Claude Fable 53.30 pts
Harvey's Legal Agent BenchmarkVals AI25.42%1of 371of 73Muse Spark 1.2—
SkillsBenchVals AI53.04%14of 2017of 35DeepSeek V4.1 Flash16.76 pts
Vals Multimodal IndexVals AI69.80%6of 176of 33Claude Fable 54.35 pts
BioMysteryBenchVals AI64.81%15of 1915of 21Claude Sonnet 5.516.30 pts
Code MigrationVals AI29.95%30of 3736of 71Claude Sonnet 5.539.88 pts
CyberBench v1.1Vals AI69.52%16of 3120of 43GPT-6 Sol8.46 pts
EMBVals AI56.98%29of 3738of 69Claude Fable 5.119.69 pts
Finance Agent (v2)Vals AI60.60%3of 373of 73Gemini 4 Argon4.80 pts
IOIVals AI21.78%30of 3135of 38Gemini 4 Argon78.22 pts
Legal Research BenchVals AI43.75%17of 3718of 72Muse Spark 1.3 Max11.54 pts
MedCodeVals AI49.35%12of 3521of 104Claude Opus 514.22 pts
MedScribeVals AI90.06%5of 365of 106Claude Opus 5.51.37 pts
Public Benefits Bench v1.1Vals AI68.47%9of 299of 45Claude Opus 58.46 pts
SAGEVals AI47.66%18of 3230of 90Claude Opus 4.78.44 pts
Tax Agent BenchVals AI56.86%32of 3738of 64Claude Fable 5.120.78 pts
DeepSWE v1.1Datacurve54.87%17of 2017of 28GPT-6 Astra19.25 pts
GDPval-AA v2.1Artificial Analysis1,481.6226of 42no list readClaude Opus 5.5364.55
AA-Briefcase v1.1Artificial Analysis1,333.3230of 41no list readClaude Opus 5.5488.53
Humanity's Last ExamArtificial Analysis45.46%20of 42no list readClaude Opus 5.515.89 pts

Nearby on Vals Index

The two models above and below on the Vals Index (Vals AI), on every column where this model has a published number.

ModelVals Index%Artificial Analysis Intelligence IndexindexTerminal-Bench 4.0%SWE-bench%MMLU Pro%MMMU Pro%Vibe Code Bench v1.1%
Kimi K3Moonshot AI50.70%4412.63%93.40%87.97%88.15%84.96%
Hunyuan Hy4 PreviewTencent49.94%–––––77.48%
Muse Spark 1.2Meta49.29%407.07%86.60%88.28%86.13%79.10%
Qwen3.8 MaxAlibaba48.27%4018.69%85.60%88.60%88.04%64.70%
DeepSeek V4 ProDeepSeek47.63%3614.14%96.40%86.97%–82.30%
Bold marks a figure better than this model’s on the same column, whichever model it belongs to. Ranks cover the models shown from the stated source.

Cost and speed, as published

Additional measurements from Artificial Analysis’ own row. Read 30 September 2026.

As publishedCost per taskOutput speed
Muse Spark 1.2 (xhigh)$0.97226.2 tok/s

Results by benchmark

Independent and vendor results retain their own harnesses and effort settings.

BenchmarkIndependent runAs reported by a publisher
τ³-BankingTool use34.85%Artificial Analysisnot published
MLCR-AAMedical reasoning31.11%Artificial Analysisnot published
EnterpriseOps-Gym-AAEnterprise workflows47.27%Artificial Analysisnot published
AutomationBench-AAWorkflow automation40.61%Artificial Analysisnot published
SciCodeScientific coding57.41%Artificial Analysisnot published
GDP.pdfDocument reasoning17.40%Artificial Analysisnot published
CritPtScientific reasoning17.71%Artificial Analysisnot published
AA-OmniscienceFactual knowledge27.2Artificial Analysisnot published
AA-LCR v1.1Long-context reasoning79.00%Artificial Analysisnot published
Terminal-Bench 4.0Agentic coding7.07%Artificial Analysis
6.06%Vals AI
not published
Terminal-Bench 2.1Agentic coding69.66%Vals AI
80.15%Artificial Analysis
not published
SWE-bench (Verified)Software engineering86.60%Vals AInot published
GPQA DiamondScience reasoning
90.40%Artificial Analysis
not published
MMLU ProKnowledge88.28%Vals AInot published
MMMU ProMultimodal reasoning86.13%Vals AInot published
Vibe Code Bench v1.1Product coding79.10%Vals AInot published
ProofBench v1.1Mathematics43.00%Vals AInot published
LegalBenchProfessional knowledge85.26%Vals AInot published
Harvey's Legal Agent BenchmarkProfessional knowledge25.42%Vals AInot published
SkillsBenchAgentic tooling53.04%Vals AInot published
Vals Multimodal IndexComposite69.80%Vals AInot published
Vals IndexComposite49.29%Vals AInot published
BioMysteryBenchAgentic science64.81%Vals AInot published
Code MigrationAgentic coding29.95%Vals AInot published
CyberBench v1.1Security69.52%Vals AInot published
EMBFinancial modelling56.98%Vals AInot published
Finance Agent (v2)Financial analysis60.60%Vals AInot published
IOICompetitive programming21.78%Vals AInot published
Legal Research BenchProfessional knowledge43.75%Vals AInot published
MedCodeHealthcare administration49.35%Vals AInot published
MedScribeHealthcare administration90.06%Vals AInot published
Public Benefits Bench v1.1Public services68.47%Vals AInot published
SAGEEducation47.66%Vals AInot published
Tax Agent BenchProfessional knowledge56.86%Vals AInot published
DeepSWE v1.1Software engineering54.87%Datacurvenot published
GDPval-AA v2.1Knowledge work1,481.62Artificial Analysisnot published
AA-Briefcase v1.1Knowledge work1,333.32Artificial Analysisnot published
Humanity's Last ExamMultidisciplinary reasoning45.46%Artificial Analysisnot published
Artificial Analysis Intelligence IndexComposite40Artificial Analysisnot published
Independent column: the suite named under each figure, one harness per suite. Publisher column: each lab’s own table, reproduced on the benchmark’s page. Publisher ranks compare reported scores in the same table row only; ties share a rank and unreported cells are excluded.

Sources

Read on 17 October 2026.