Apple ML Research: Muse Spark Open with MLX on Mac Studio
The M5 Ultra, compared to the M3 Ultra, demonstrates significant speedups in Time to First Token (TTFT) for Qwen 3.8 and Mistral Medium 3.5, especially in compute-bound workloads. Its 1.2 TB/s unified memory bandwidth (vs 819 GB/s on M3 Ultra) also provides a key advantage in memory-bandwidth bound decode operations. The Mac Studio with M5 Ultra consumes about 300W, making it energy-efficient for LLM inference and suitable for home or office AI deployment.
This report uniquely details the M5 Ultra's performance gains over the M3 Ultra, specifically highlighting its 1.2 TB/s memory bandwidth compared to the M3 Ultra's 819 GB/s.
Time & source
Times shown in UTC
Display time zone: UTC
Local time zone unavailable; showing UTC.
IngestedOffset at this time: UTC+0Oct 16, 2026, 15:00 UTC
- Ingested
- Oct 16, 2026, 15:00
- Source type
- Unclassified
- Basis
- Running about 3.7× the median of this source's recent listed items
- Triggering item
- Apple ML Research: Muse Spark Open with MLX on Mac Studio
- Metric comparison
- 287 vs median 78 (20 baseline samples)
- Detected
- 10/16, 23:00
Full text isn't available here.
Read at source →