Meta released Llama 3.1 in 8B, 70B and 405B sizes with a 128K context window on 23 July 2024 [1], Llama 3.2 with 1B and 3B phone models and 11B and 90B vision models at Connect 2024 [2], and Llama 3.3 70B on 6 December [3]. Muse Spark, released on 8 April 2026, was the first model from Meta Superintelligence Labs [4]. Muse Spark 1.1 added a 1M-token context window [5], 1.2 was a coding-focused update [6] and 1.3, on 2 September, was the fourth Muse Spark model in five months [7]. Muse Wonder and Muse Wonder Full, built around 'Active Memory', followed on 12 October [8].