Meta's Muse Spark 1.1 evaluation report says the model scored 59% on CyberGym vulnerability tasks, and posts on X quickly framed that result as a clear improvement over Muse Spark 1.0's 43.5% while still trailing several rivals. In one widely shared post on X, researcher @scaling01 wrote that Muse Spark 1.1 was "worse than Opus 4.6 and Sonnet 4.6, but better than Opus 4.5," turning a dry benchmark table into a pecking-order argument on X.