evals.report
BenchmarksLabsCompareRun guidesIn the wild
MetaMuse

Muse Spark

Meta · Muse. Released Apr 8, 2026.

Muse Spark is a model from Meta in the Muse family, released Apr 8, 2026. evals.report tracks 21 reported Muse Spark benchmark scores across LMArena, SWE-bench Pro, GPQA Diamond, MMMU-Pro, AIME (OTIS Mock), SimpleQA Verified, FrontierMath, Artificial Analysis Intelligence Index, and 13 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

21 results

Benchmark results 21

Compare this model
BenchmarkCategoryScoreMetricStatusDate
LMArenaChat preference1474source-defined ratingOfficialApr 8, 2026Details
SWE-bench ProCoding55.00%% resolvedOfficialApr 8, 2026Details
GPQA DiamondReasoning89.8%accuracyOfficialApr 8, 2026Details
MMMU-ProMultimodal80.4%accuracyOfficialApr 8, 2026Details
AIME (OTIS Mock)Reasoning88.9%accuracyOfficialApr 8, 2026Details
SimpleQA VerifiedOther66.3%accuracyOfficialApr 8, 2026Details
FrontierMathReasoning39.0%accuracyOfficialApr 8, 2026Details
Artificial Analysis Intelligence IndexReasoning52.1IndexUnverifiedApr 8, 2026Details
Epoch Capabilities IndexReasoning155.1IndexOfficialApr 8, 2026Details
τ²-bench (Telecom)Tool use91.5%pass^1OfficialApr 8, 2026Details
GDPvalAgents1417EloOfficialApr 8, 2026Details
AA-Omniscience: Knowledge and Hallucination BenchmarkReasoning4AA-Omniscience IndexOfficialApr 8, 2026Details
IFBenchReasoning75.9%accuracyOfficialApr 8, 2026Details
MultiChallengeReasoning75.52%accuracyVerifiedApr 8, 2026Details
WebDev ArenaChat preference1508EloVerifiedApr 8, 2026Details
Design ArenaChat preference1306EloVerifiedApr 8, 2026Details
CharXivMultimodal86.4%accuracyUnverifiedApr 8, 2026Details
ScreenSpot-ProMultimodal84.1%accuracyUnverifiedApr 8, 2026Details
FrontierMath Tier 4Reasoning14.6%accuracyOfficialApr 8, 2026Details
Vibe Code BenchCoding19.67%Overall accuracyVerifiedApr 8, 2026Details
MultiNRCReasoning59.05%accuracyOfficialApr 8, 2026Details