evals.report
BenchmarksLabsCompareRun guidesIn the wild
MetaMuse

Muse Spark 1.1

Meta · Muse. Released Jul 9, 2026.

Muse Spark 1.1 is a model from Meta in the Muse family, released Jul 9, 2026. evals.report tracks 6 reported Muse Spark 1.1 benchmark scores across SWE-bench Pro, Terminal-Bench 2.1, OSWorld, Humanity's Last Exam, OpenAI-MRCR v2 (Multi-Round Coreference Resolution), MCP Atlas — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

6 results

Benchmark results 6

Compare this model
BenchmarkCategoryScoreMetricStatusDate
SWE-bench ProCoding61.5%% resolvedOfficialJul 9, 2026Details
Terminal-Bench 2.1Agents80.0%task successVerifiedJul 9, 2026Details
OSWorldAgents80.8%task success rateVerifiedJul 9, 2026Details
Humanity's Last ExamReasoning52.2%accuracyVerifiedJul 9, 2026Details
OpenAI-MRCR v2 (Multi-Round Coreference Resolution)Reasoning54.1%accuracy (mean SequenceMatcher similarity)VerifiedJul 9, 2026Details
MCP AtlasTool use88.1%pass rateVerifiedJul 9, 2026Details