Is the RTX 5090 faster than the M5 Max for LLMs?

The RTX 5090 reads memory about 192% faster than the M5 Max (1,792 vs 614 GB/s), so it writes about 192% faster with any model that fits both. In llama.cpp’s measured test the RTX 5090 reads a prompt about 4.4× as fast (14,073 vs 3,220 tokens a second). Reading a prompt depends on GPU compute rather than memory bandwidth, so this sets the wait before the first word with a long document or codebase.

Which can run bigger models, the M5 Max or the RTX 5090?

The M5 Max has about 96 GB for models, the RTX 5090 about 31 GB. 47 of 73 popular open models run well on the M5 Max, 38 on the RTX 5090.

Can MLX Node run on the RTX 5090?

No. MLX Node and MLX Harness run on Apple Silicon Macs only, with Apple’s MLX framework.

M5 Max vs RTX 5090 for local LLMs: memory, speed, which to buy

Side by side

	M5 Max	RTX 5090
What it is	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)	A graphics card for a desktop PC
Memory sizes	36, 48, 64, 128 GB	32 GB
Memory bandwidth	460–614 GB/s	1,792 GB/s
For models, at most	about 96 GB	about 31 GB
Models that run well	47 of 73	38 of 73
Measured: writing	120 tok/s	290 tok/s
Measured: reading a prompt	3,220 tok/s	14,073 tok/s

Measured rows: each machine’s fastest result in llama.cpp’s public benchmarks with Llama 2 7B Q4_0, writing 128 tokens (tg128) and reading a 512-token prompt (pp512), in tokens a second. The Mac ran Metal, the RTX 5090 CUDA. Apple Silicon results · RTX 5090 results For models we count the RTX 5090’s own 32 GB less about 1 GB, and about 75% of the Mac’s unified memory, the macOS default.

Where each one wins

Writing speed. The RTX 5090 reads memory about 192% faster than the M5 Max (1,792 vs 614 GB/s), so it writes about 192% faster with any model that fits both.
Model size. The M5 Max has about 96 GB for models, the RTX 5090 about 31 GB. 47 of 73 popular open models run well on the M5 Max, 38 on the RTX 5090.
Reading long prompts. In llama.cpp’s measured test the RTX 5090 reads a prompt about 4.4× as fast (14,073 vs 3,220 tokens a second). Reading a prompt depends on GPU compute rather than memory bandwidth, so this sets the wait before the first word with a long document or codebase.
Software. The RTX 5090 runs NVIDIA’s CUDA, which most AI software supports first, including training code and many image and video tools. On a Mac you run models with MLX, llama.cpp’s Metal backend or PyTorch’s MPS backend instead.
Around it. A graphics card needs a PC around it, and its own 32 GB is the limit: a model that doesn’t fit has to run partly from the PC’s much slower system memory, or across a second card.
The Mac. It’s also an everyday computer, and it runs MLX Harness and MLX Node, which are made for Apple Silicon.

Speed with popular models

Tokens a second for one conversation, each chip at its largest memory size (128 GB and 32 GB), with the best version of the model that fits.

Model	M5 Max	RTX 5090
Llama 3.1 8B Instruct8B	33–58 8-bit	95–169 8-bit
gpt-oss-20b21B · 3.6B active	145–259 native 4.25-bit	424–755 native 4.25-bit
Qwen3.6 35B A3B35B · 3B active	86–154 8-bit	336–597 6-bit
Qwen3.8 27B27B	10–17 8-bit	28–50 8-bit
Llama 3.3 70B Instruct70B	4–7 8-bit	37–65 2-bit lower quality
gpt-oss-120b117B · 5.1B active	102–182 native 4.25-bit	too big
Qwen3.5 122B A10B122B · 10B active	49–88 4-bit	too big
DeepSeek V4 Flash 0731284B · 13B active	67–120 2-bit lower quality	too big

Questions

Is the RTX 5090 faster than the M5 Max for LLMs?: The RTX 5090 reads memory about 192% faster than the M5 Max (1,792 vs 614 GB/s), so it writes about 192% faster with any model that fits both. In llama.cpp’s measured test the RTX 5090 reads a prompt about 4.4× as fast (14,073 vs 3,220 tokens a second). Reading a prompt depends on GPU compute rather than memory bandwidth, so this sets the wait before the first word with a long document or codebase.
Which can run bigger models, the M5 Max or the RTX 5090?: The M5 Max has about 96 GB for models, the RTX 5090 about 31 GB. 47 of 73 popular open models run well on the M5 Max, 38 on the RTX 5090.
Can MLX Node run on the RTX 5090?: No. MLX Node and MLX Harness run on Apple Silicon Macs only, with Apple’s MLX framework.

More comparisons

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.

RTX 5090 figures: NVIDIA: GeForce RTX 5090 (specs: 32 GB GDDR7, 512-bit, 575 W TGP), NVIDIA RTX Blackwell GPU Architecture whitepaper (Tables 1 and 4).