Question 1

Which Mac do I need for gpt-oss 120B?

Accepted Answer

More than 64 GB of memory: it needs about 66 GB and 64 GB Macs let the GPU use 56 GB. The cheapest is the Mac Studio M5 Max with 128 GB ($5,099), at about 89 tokens per second. The MacBook Pro M5 Max with 128 GB is as fast. The Mac Studio M5 Ultra with 96 GB ($5,499) writes about 113.

Question 2

How much memory does a 70B model need?

Accepted Answer

Llama 3.3 70B in 4-bit needs about 40 GB of weights, plus 11 GB of KV cache for a 32,768-token context. A 64 GB Mac runs it, slowly: about 6 tokens per second on an M5 Pro, 13 on an M5 Max. The M5 Ultra (96 GB at least) writes about 23.

Question 3

Can a MacBook Neo or a MacBook Air run a local LLM?

Accepted Answer

Yes, small ones. The MacBook Neo has 8 GB: models up to 4B parameters, like Qwen3.5 4B at about 19 tokens per second. A 16 GB MacBook Air runs models up to 14B (Phi-4 at about 15 tokens per second, Llama 3.1 8B at about 28). They have no fan: long prompts slow down as the chip heats up.

Question 4

MLX or llama.cpp (Ollama, LM Studio)?

Accepted Answer

On a Mac, MLX is usually a little faster. It is much faster on hybrid models like Qwen3.5 and Qwen3.6 (1.4 to 1.9 times), and it reads prompts faster on M5 and M6. LM Studio runs both. Ollama runs some models on MLX and the others on llama.cpp.

Question 5

Why does a 35B model run faster than an 8B one?

Accepted Answer

Because it is a mixture of experts. Qwen3.6 35B A3B only uses 3 billion parameters per token: with MLX it writes faster than a dense 8B model, but its 35 billion parameters must all fit in memory.

Question 6

Can I run DeepSeek or Kimi on a Mac?

Accepted Answer

DeepSeek V4 Flash, yes: in 4-bit it needs about 150 GB and runs on a 256 GB Mac Studio M5 Ultra, at about 48 tokens per second. DeepSeek V3.2 and Kimi need the 512 GB Mac Studio M5 Ultra, which ships in late October 2026. DeepSeek V3.2 in 4-bit needs about 380 GB and writes about 25 tokens per second. Kimi K2.6 (1,000 billion parameters) only fits in 3-bit.

Question 7

Are my settings sent anywhere?

Accepted Answer

No. The math runs in your browser. Nothing is sent until you click Save or Share.

Parameter	flash (active)	standard
Mac	Mac Studio	Mac Studio
Chip	M5 Ultra, 36-core CPU, 80-core GPU	M5 Ultra, 36-core CPU, 80-core GPU
Unified memory	256 GB	512 GB
Model	GLM-5.3-Flash	GLM-5.3
Context	131,072 tokens	131,072 tokens

Result	flash (active)	standard
Writing speed	42 tok/s	22 tok/s
Price	$10,799	–
Writing at full context	40 tok/s	21 tok/s
Reading speed	875 tok/s	462 tok/s
Time to read the context	2 min 31 s	4 min 52 s
Memory needed	207 GB	432 GB
Memory usable by the GPU	239 GB	498 GB
Models that fit	35	38
Memory bandwidth	1,228 GB/s	1,228 GB/s

Which Mac for local LLMs? Models and speed (tok/s) for every Mac

Scenarios

Results