Is the M4 Max faster than the M2 Ultra for LLMs?

The M2 Ultra reads memory about 47% faster than the M4 Max (800 vs 546 GB/s), so it writes about 47% 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 M2 Ultra or the M4 Max?

The M2 Ultra: The M2 Ultra goes up to 192 GB of memory, the M4 Max to 128 GB. At its largest it runs 49 of 73 open models well, against 47.

M2 Ultra vs M4 Max for local LLMs: memory, speed, which to buy

Side by side

	M2 Ultra	M4 Max
Memory sizes	64, 128, 192 GB	36, 48, 64, 128 GB
Memory bandwidth	800 GB/s	410–546 GB/s
For models, at most	about 144 GB	about 96 GB
Models that run well	49 of 73	47 of 73
MLX seat class	Ultra	Pro
Found in	Mac Studio (2023), Mac Pro (2023), Mac Pro (Rack, 2023)	MacBook Pro (14-inch, M4 Pro or M4 Max, 2024), MacBook Pro (16-inch, 2024), Mac Studio (2025)

Speed with popular models

Tokens a second for one conversation, each chip at its largest memory size (192 GB and 128 GB), with the best version of the model that fits.

Model	M2 Ultra	M4 Max
Llama 3.1 8B Instruct8B	42–75 8-bit	29–51 8-bit
gpt-oss-20b21B · 3.6B active	189–337 native 4.25-bit	129–230 native 4.25-bit
Qwen3.6 35B A3B35B · 3B active	113–200 8-bit	77–137 8-bit
Qwen3.8 27B27B	13–22 8-bit	11–20 6-bit
Llama 3.3 70B Instruct70B	5–9 8-bit	3–6 8-bit
gpt-oss-120b117B · 5.1B active	133–237 native 4.25-bit	91–162 native 4.25-bit
Qwen3.5 122B A10B122B · 10B active	34–60 8-bit	44–78 4-bit
DeepSeek V4 Flash 0731284B · 13B active	63–112 3-bit lower quality	60–107 2-bit lower quality

Questions

Is the M4 Max faster than the M2 Ultra for LLMs?: The M2 Ultra reads memory about 47% faster than the M4 Max (800 vs 546 GB/s), so it writes about 47% 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 M2 Ultra or the M4 Max?: The M2 Ultra: The M2 Ultra goes up to 192 GB of memory, the M4 Max to 128 GB. At its largest it runs 49 of 73 open models well, against 47.

More comparisons

The M2 Ultra for local AI · The M4 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.