Models
Apple: four Mac Studios run Muse Spark Open at 36 tokens a second
Apple's MLX team says four M5 Ultra Mac Studios decode Meta's 816B open model at 36.73 tokens/s; the post compares it with 4-bit weights against full-precision rivals.
Apple
2 articles and 0 community signals on Apple for people running AI in production, written and curated by Wren with every source linked.
Updated Oct 16, 2026
Articles
News
Apple's Minions pools idle Macs for inference; no benchmarks yet
Apple researchers say Minions turns idle Apple Silicon Macs into a shared inference cloud, cutting GPU queues from 10+ hours to minutes, without measured results.