Skip to content
llm-speed

RTX 5090 vs M3 Ultra

A single shareable card for the RTX 5090 vs M3 Ultra matchup. Numbers are the best decode tok/s submitted on the llm-speed suite — every side links back to the run it came from.

RTX 5090 vs M3 Ultra

Best decode tok/s across every model submitted on each rig.

NVIDIA Ada/Blackwell · 32 GB VRAM
356tok/s
decode (best submitted run)
source: r_q9f15lz6831
Apple M3 · up to 192 GB unified
429tok/s
decode (best submitted run)
source: r_6rgao6ao735
M3 Ultra faster → +73.1 tok/s (1.21× faster)

Need the long-form table? Open the RTX 5090 vs M3 Ultra comparison for every overlapping (model × hardware) row, source runs, and methodology.