Model ReleasesLiquid AI breaks the efficiency barrier with LFM2.5-24B-A2B, a 24B MoE model that delivers elite reasoning with only 2B active parameters for local execution.3 дня назад5 мин чтения
