Is the M3 Max faster than the M2 Max for LLMs?

About the same: their memory bandwidth is within a few percent. The M3 Max goes up to 128 GB of memory, the M2 Max to 96 GB.

Which can run bigger models, the M2 Max or the M3 Max?

The M3 Max: The M3 Max goes up to 128 GB of memory, the M2 Max to 96 GB. At its largest it runs 47 of 73 open models well, against 43.

M2 Max vs M3 Max for local LLMs: memory, speed, which to buy

Side by side

	M2 Max	M3 Max
Memory sizes	32, 64, 96 GB	36, 48, 64, 96, 128 GB
Memory bandwidth	400 GB/s	300–400 GB/s
For models, at most	about 72 GB	about 96 GB
Models that run well	43 of 73	47 of 73
MLX seat class	Pro	Pro
Found in	MacBook Pro (14-inch, 2023), MacBook Pro (16-inch, 2023), Mac Studio (2023)	MacBook Pro (14-inch, M3 Pro or M3 Max, Nov 2023), MacBook Pro (16-inch, Nov 2023)

Speed with popular models

Tokens a second for one conversation, each chip at its largest memory size (96 GB and 128 GB), with the best version of the model that fits.

Model	M2 Max	M3 Max
Llama 3.1 8B Instruct8B	21–38 8-bit	21–38 8-bit
gpt-oss-20b21B · 3.6B active	95–168 native 4.25-bit	95–168 native 4.25-bit
Qwen3.6 35B A3B35B · 3B active	56–100 8-bit	56–100 8-bit
Qwen3.8 27B27B	12–21 4-bit	12–21 4-bit
Llama 3.3 70B Instruct70B	3–6 6-bit	2–4 8-bit
gpt-oss-120b117B · 5.1B active	67–119 native 4.25-bit	67–119 native 4.25-bit
Qwen3.5 122B A10B122B · 10B active	41–73 3-bit lower quality	32–57 4-bit
DeepSeek V4 Flash 0731284B · 13B active	too big	44–78 2-bit lower quality

Questions

Is the M3 Max faster than the M2 Max for LLMs?: About the same: their memory bandwidth is within a few percent. The M3 Max goes up to 128 GB of memory, the M2 Max to 96 GB.
Which can run bigger models, the M2 Max or the M3 Max?: The M3 Max: The M3 Max goes up to 128 GB of memory, the M2 Max to 96 GB. At its largest it runs 47 of 73 open models well, against 43.

More comparisons

The M2 Max for local AI · The M3 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.