Is the M5 Max faster than the M3 Ultra for LLMs?

The M3 Ultra reads memory about 33% faster than the M5 Max (819 vs 614 GB/s), so it writes about 33% faster with any model that fits both. Reading a long prompt depends on GPU compute instead, which this comparison doesn’t cover.

Which can run bigger models, the M3 Ultra or the M5 Max?

The M3 Ultra: The M3 Ultra goes up to 512 GB of memory, the M5 Max to 128 GB. At its largest it runs 60 of 73 open models well, against 47.

M3 Ultra vs M5 Max for local LLMs: memory, speed, which to buy

Side by side

	M3 Ultra	M5 Max
Memory sizes	96, 256, 512 GB	36, 48, 64, 128 GB
Memory bandwidth	819 GB/s	460–614 GB/s
For models, at most	about 384 GB	about 96 GB
Models that run well	60 of 73	47 of 73
MLX seat class	Ultra	Pro Latest
Found in	Mac Studio (2025)	MacBook Pro (14-inch, M5 Pro or M5 Max), MacBook Pro (16-inch, M5 Pro or M5 Max), Mac Studio (M5 Max or M5 Ultra)

Speed with popular models

Tokens a second for one conversation, each chip at its largest memory size (512 GB and 128 GB), with the best version of the model that fits.

Model	M3 Ultra	M5 Max
Llama 3.1 8B Instruct8B	43–77 8-bit	33–58 8-bit
gpt-oss-20b21B · 3.6B active	194–345 native 4.25-bit	145–259 native 4.25-bit
Qwen3.6 35B A3B35B · 3B active	115–205 8-bit	86–154 8-bit
Qwen3.8 27B27B	13–23 8-bit	10–17 8-bit
Llama 3.3 70B Instruct70B	5–9 8-bit	4–7 8-bit
gpt-oss-120b117B · 5.1B active	137–243 native 4.25-bit	102–182 native 4.25-bit
Qwen3.5 122B A10B122B · 10B active	35–62 8-bit	49–88 4-bit
DeepSeek V4 Flash 0731284B · 13B active	27–47 8-bit	67–120 2-bit lower quality

Questions

Is the M5 Max faster than the M3 Ultra for LLMs?: The M3 Ultra reads memory about 33% faster than the M5 Max (819 vs 614 GB/s), so it writes about 33% faster with any model that fits both. Reading a long prompt depends on GPU compute instead, which this comparison doesn’t cover.
Which can run bigger models, the M3 Ultra or the M5 Max?: The M3 Ultra: The M3 Ultra goes up to 512 GB of memory, the M5 Max to 128 GB. At its largest it runs 60 of 73 open models well, against 47.

More comparisons

The M3 Ultra for local AI · The M5 Max for local AI · Compare every model · The best Mac for local LLMs

How it’s worked out

Memory: the model’s parameters at 8, 6, 4, 3 or 2 bits per weight plus MLX’s group scales (half a bit per weight), or its native format, with room for the context on top. By default macOS lets the GPU use about 75% of unified memory. Speed: memory bandwidth divided by the bytes read per token is the ceiling for one conversation; in llama.cpp’s public Apple Silicon benchmarks real runs reach about 45% to 80% of it, which is the range shown. These are estimates, not measurements: measure your own Mac on MLX Bench. Model figures from each official model card, chip figures from Apple’s Tech Specs. Checked 2026-10-19. Same data as mlx.app/compare.