Meta's Muse Spark 1.1 is getting attention on X through posts pointing back to a Vals AI announcement post that claims state-of-the-art results on TaxEval, MedScribe and Harvey's Legal Agent Bench. In one widely shared reaction, @shuchaobi wrote on X, "We are #1 on Harvey’s Legal Agent Bench," while @danielgross posted, "SOTA at TaxEval!" The performance and cost comparisons are being presented as claims from those posts, not as independently verified results.