evals.report
BenchmarksLabsCompareRun guidesIn the wild
MetaMuse

Muse Glimmer 30B

Meta · Muse. Released Aug 10, 2026.

Muse Glimmer 30B is a model from Meta in the Muse family, released Aug 10, 2026. evals.report tracks 15 reported Muse Glimmer 30B benchmark scores across AA-Omniscience v1.1: Knowledge and Hallucination Benchmark, AA-Omniscience: Knowledge and Hallucination Benchmark, MCP Atlas, SWE-bench Pro, SWE-bench Verified, Terminal-Bench 2.1, SciCode, CharXiv, and 7 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

Open15 results

Benchmark results 15

Compare this model
BenchmarkCategoryScoreMetricStatusDate
AA-Omniscience v1.1: Knowledge and Hallucination BenchmarkReasoning-32AA-Omniscience Net ScoreOfficialAug 10, 2026Details
AA-Omniscience: Knowledge and Hallucination BenchmarkReasoning-33AA-Omniscience IndexOfficialAug 10, 2026Details
MCP AtlasTool use75.5%pass rateVerifiedAug 10, 2026Details
SWE-bench ProCoding51.2%% resolvedVerifiedAug 10, 2026Details
SWE-bench VerifiedCoding76.0%% resolvedVerifiedAug 10, 2026Details
Terminal-Bench 2.1Agents51.7%task successVerifiedAug 10, 2026Details
SciCodeCoding43.6%accuracyVerifiedAug 10, 2026Details
CharXivMultimodal78.8%accuracyVerifiedAug 10, 2026Details
ScreenSpot-ProMultimodal75.4%accuracyVerifiedAug 10, 2026Details
MMMU-ProMultimodal74%accuracyVerifiedAug 10, 2026Details
OSWorldAgents65.9%task success rateVerifiedAug 10, 2026Details
IFBenchReasoning77.0%accuracyVerifiedAug 10, 2026Details
GPQA DiamondReasoning83.5%accuracyVerifiedAug 10, 2026Details
Humanity's Last ExamReasoning22.0%accuracyVerifiedAug 10, 2026Details
AIME 2026Reasoning94.7%accuracyVerifiedAug 10, 2026Details