Apple · M3 Ultra · Desktop

Best AI Models for Mac Studio (M3 Ultra, 256GB)

Runs 27–31B and 120B MoE models well; dense 70B fits but is slow at 12 tok/s.

256 GB unified252.5 GB for models819 GB/s66 TFLOPS FP16270 W2025

Memory

96 GB$3,999 256 GB$7,099 512 GB$9,499

What LLMs can the Mac Studio (M3 Ultra, 256GB) run?

#	Model	Speed	Quant · Memory
1	MiMo V2.6 Flash RLXiaomiHow to run → Q4_K_M · 186.5 GB · 39.5 t/s	39.5 tok/s	Q4_K_M · 186.5 GB
2	DeepSeek V4 Flash 0731DeepSeek Q4_K_M · 183.0 GB · 26.8 t/s	26.8 tok/s	Q4_K_M · 183.0 GB
3	GLM 5.3 FlashZ.ai Q4_K_M · 205.2 GB · 35.3 t/s	35.3 tok/s	Q4_K_M · 205.2 GB
4	Qwen3.8 27BAlibaba Q4_K_M · 17.9 GB · 32.9 t/s	32.9 tok/s	Q4_K_M · 17.9 GB
5	Hy3Tencent Q4_K_M · 182.3 GB · 26.3 t/s	26.3 tok/s	Q4_K_M · 182.3 GB

#	Model	Speed	Quant · Memory
1	MiMo V2.6 Flash RLXiaomiHow to run → Q4_K_M · 186.5 GB · 39.5 t/s	39.5 tok/s	Q4_K_M · 186.5 GB
2	DeepSeek V4 Flash 0731DeepSeek Q4_K_M · 183.0 GB · 26.8 t/s	26.8 tok/s	Q4_K_M · 183.0 GB
3	Qwen3.8 27BAlibaba Q4_K_M · 17.9 GB · 32.9 t/s	32.9 tok/s	Q4_K_M · 17.9 GB
4	GLM 5.3 FlashZ.ai Q4_K_M · 205.2 GB · 35.3 t/s	35.3 tok/s	Q4_K_M · 205.2 GB
5	Hy3Tencent Q4_K_M · 182.3 GB · 26.3 t/s	26.3 tok/s	Q4_K_M · 182.3 GB

#	Model	Speed	Quant · Memory
1	DeepSeek V4 Flash 0731DeepSeekHow to run → Q4_K_M · 183.0 GB · 26.8 t/s	26.8 tok/s	Q4_K_M · 183.0 GB
2	GLM 5.3 FlashZ.ai Q4_K_M · 205.2 GB · 35.3 t/s	35.3 tok/s	Q4_K_M · 205.2 GB
3	Qwen3.8 27BAlibaba Q4_K_M · 17.9 GB · 32.9 t/s	32.9 tok/s	Q4_K_M · 17.9 GB
4	Inkling SmallThinking Machines Q4_K_M · 162.7 GB · 98.0 t/s	98.0 tok/s	Q4_K_M · 162.7 GB
5	Hy3Tencent Q4_K_M · 182.3 GB · 26.3 t/s	26.3 tok/s	Q4_K_M · 182.3 GB

#	Model	Speed ↓	Quant · Memory
1	Qwen3.5 4BAlibabaHow to run → Q4_K_M · 3.4 GB · 176 t/s	176 tok/s	Q4_K_M · 3.4 GB
2	GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 107 t/s	107 tok/s	Q4_K_M · 13.1 GB
3	Inkling SmallThinking Machines Q4_K_M · 162.7 GB · 98.0 t/s	98.0 tok/s	Q4_K_M · 162.7 GB
4	Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 92.7 t/s	92.7 tok/s	Q4_K_M · 16.5 GB
5	Hy3Tencent Q4_K_M · 182.3 GB · 26.3 t/s	26.3 tok/s	Q4_K_M · 182.3 GB

#	Model	Speed	Quant · Memory
1	GLM 5.3 FlashZ.aiHow to run → Q4_K_M · 205.2 GB · 35.3 t/s	35.3 tok/s	Q4_K_M · 205.2 GB
2	Qwen3.8 27BAlibaba Q4_K_M · 17.9 GB · 32.9 t/s	32.9 tok/s	Q4_K_M · 17.9 GB
3	Inkling SmallThinking Machines Q4_K_M · 162.7 GB · 98.0 t/s	98.0 tok/s	Q4_K_M · 162.7 GB
4	Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 32.9 t/s	32.9 tok/s	Q4_K_M · 17.9 GB
5	NVIDIA Nemotron 3.5 Super VL 120B A12B BF16NVIDIA Q4_K_M · 87.7 GB · 42.7 t/s	42.7 tok/s	Q4_K_M · 87.7 GB

#	Model	Speed	Quant · Memory
1	GLM 5.3 FlashZ.aiHow to run → Q4_K_M · 205.2 GB · 35.3 t/s	35.3 tok/s	Q4_K_M · 205.2 GB
2	Qwen3.8 27BAlibaba Q4_K_M · 17.9 GB · 32.9 t/s	32.9 tok/s	Q4_K_M · 17.9 GB
3	MiniMax M2.7MiniMax Q4_K_M · 139.6 GB · 41.7 t/s	41.7 tok/s	Q4_K_M · 139.6 GB
4	GLM 4.5Z.ai Q4_K_M · 217.9 GB · 17.9 t/s	17.9 tok/s	Q4_K_M · 217.9 GB
5	Hy3Tencent Q4_K_M · 182.3 GB · 26.3 t/s	26.3 tok/s	Q4_K_M · 182.3 GB

See all 425 models on the Mac Studio (M3 Ultra, 256GB) →Bars = traction for models without a Score yet

About the Mac Studio (M3 Ultra, 256GB)

The Mac Studio (M3 Ultra, 256GB) runs 27–31B and 120B MoE models well. macOS keeps about 3.5 GB, leaving 252.5 GB for models. Prompt reading is its weak spot: a 4K-token prompt waits 9.4 s before the first word on MiMo V2.6 Flash RL. At its $7,099 MSRP it gives 10.8 tok/s per $1,000, 5th of the 6 machines compared here. M3 Ultra generation, released 2025.

For models252.5 GBmacOS keeps 3.5 GB

ChipM3 Ultrareleased 2025

Mac Studio (M3 Ultra, 256GB) strengths and weaknesses

Strengths

Runs 120B MoE models: GPT OSS 120B at 75.5 tok/s in 70.8 GB
Room for long context: Qwen3.8 27B at Q4 uses 17.9 GB, leaving 234.6 GB free

Weaknesses

Slow prompt reading: A 32K-token prompt on MiMo V2.6 Flash RL waits 1.3 min for the first token
No CUDA: vLLM, AWQ, GPTQ and FP8 builds don't run; use GGUF or MLX

Memory for models

252.5 GB of 256 GB

35B MoE22 GB70B51.3 GB120B MoE70.8 GB427B242 GB

120B MoE70.8 GB427B242 GB

For models 252.5 GBmacOS keeps 3.5 GBModel size at Q4

Everything on the size ladder fits. MiMo V2.6 Flash RL and Qwen3.8 27B together use 204.4 GB and leave 48.1 GB free.

Speed per dollar and per watt

NewUsed

On Qwen3.6 35B A3B (35B params, 3B active) at Q4_K_M. New means a shop price, either a street quote or the maker's own price; used is compared with used. The last two columns are the speed you get per $1,000 of price and per 100 W of power; higher is better. The Mac Studio (M3 Ultra, 256GB) has no current new price, so it is shown at its MSRP.

Machine	New price	tok/s	tok/s per $1,000	tok/s per 100 W
RTX 509032 GB	about $5,376	213	39.7	37.1
Mac Studio M3 Ultra96 GB	about $3,999	76.6	19.2	28.4
Mac Studio M4 Max128 GB	about $3,499	64.2	18.3	44.3
Mac Studio M2 Ultra192 GB	about $5,599	75.9	13.6	25.7
Mac Studio M3 Ultra256 GB	about $7,099	76.6	10.8	28.4
Mac Studio M3 Ultra512 GB	about $9,499	76.6	8.1	28.4

Machine	Used price	tok/s	tok/s per $1,000	tok/s per 100 W
RTX 409024 GBBest used value	about $3,000	151	50.3	33.6
RTX 509032 GB	about $6,400	213	33.3	37.1
Mac Studio M3 Ultra96 GB	about $4,525	76.6	16.9	28.4

Cost per token: the Mac Studio (M3 Ultra, 256GB) against 24–512 GB machines and GPUs

On Qwen3.6 35B A3B (35B params, 3B active, Q4_K_M, 4K context). Speed, cost per token and power per token depend on the model.

24–36 GB96–128 GB192+ GBthis desktopat MSRP

Cost per 1M output tokenson Qwen3.6 35B A3B · card over 3 years + power

$6.02

Machine, MSRP: $5.88
Power: $0.15

3-year cost of ownership$7,276

Include card price

Power$/kWhUseh/dayKeepyears

Each dot is a machine or GPU; hover or tap it for its numbers, click to open it. Machines without a price for the chosen basis are left out. How we calculate cost per token

Similar devices and alternatives

Other memory sizes of the same machine first, then machines and GPUs people weigh against it, each with the one reason it is worth a look. Speeds on Qwen3.6 35B A3B.

Same chip, 160 GB less memory for $3,100 lessMac Studio M3 Ultra (96 GB)Apple · 92.5 GB for models · M3 Ultra

same speed▼ $3,100 less▼ 160 GB$3,999

Same family Same chip, 2× the memory for $2,400 moreMac Studio M3 Ultra (512 GB)Apple · 508.5 GB for models · M3 Ultra

same speed▲ $2,400 more▲ 256 GB$9,499

Same family About as fastMac Studio M2 Ultra (192 GB)Apple · 188.5 GB for models · M2 Ultra

same speed▼ 64 GB$5,599

Step down to 124.5 GBMac Studio M4 Max (128 GB)Apple · 124.5 GB for models · M4 Max

▼ 16% speed▼ 128 GB$3,499

2.0× faster if your models fit in 24 GBRTX 4090NVIDIA · 24 GB VRAM · GPU

▲ 97% speed▼ 228.5 GB$1,599rent $0.34/hr

2.8× faster if your models fit in 32 GBRTX 5090NVIDIA · 32 GB VRAM · GPU

▲ 179% speed▼ 220.5 GB$5,376rent $0.50/hr

Mac Studio (M3 Ultra, 256GB) tokens per second by model size

Output speed depends on memory bandwidth (819 GB/s). Grey rows need more than the 252.5 GB available.

9BQwen3.5 9B · Q4_K_M · 6.6 GB
90.1tok/s
14BQwen3 14B · Q4_K_M · 11.9 GB
60.3tok/s
27–31BQwen3.8 27B · Q4_K_M · 17.9 GB
32.9tok/s
35B MoEQwen3.6 35B A3B · Q4_K_M · 22.0 GB
76.6tok/s
70BLlama 3.3 70B Instruct · Q4_K_M · 51.3 GB
12.3tok/s
120B MoEGPT OSS 120B · Q4_K_M · 70.8 GB
75.5tok/s

Estimated speeds: Q4_K_M, 4K context, one user. How we estimate

Ultra generations: writing vs compute

M1 Ultra → M3 Ultra

Writing speed (tok/s)Compute (TFLOPS FP16)

Writing speed follows memory bandwidth and barely moved: 75.9 to 76.6 tok/s since the M1 Ultra. Compute, which sets how fast prompts are read, grew 2.1×. Indexed to the M1 Ultra; speeds on Qwen3.6 35B A3B, same 128 GB.

All 1329 models on the Mac Studio (M3 Ultra, 256GB)

Featured only

Best quant that fits in 252.5 GB, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works

LLM models compatible with Mac Studio (M3 Ultra, 256GB) — ranked by Score, recency and Fit
Model	Quant	VRAM	Speed	Context	Fit
Jan Code 4B4.4BChatFunctionsCode Q4_K_M·182.6 t/s	Q4_K_M	4.5 GB2%	182.6 t/s	262K	S99, Runs great
Helium 1 2B2.0BChat Q4_K_M·327.6 t/s	Q4_K_M	2.0 GB1%	327.6 t/s	4K	S100, Runs great
Llama 3.1 70B LatamGPT SFT 1.070.6BChat Q4_K_M·13.2 t/s	Q4_K_M	44.0 GB17%	13.2 t/s	4K	B52, Decent
Bitnet B1.58 2B 4T850MChat BF16·265.4 t/s	BF16	2.3 GB1%	265.4 t/s	4K	S100, Runs great
OpenThinker3 1.5B1.5BChat Q4_K_M·447.9 t/s	Q4_K_M	1.7 GB1%	447.9 t/s	33K	S100, Runs great
MN VelvetCafe RP 12B12.2BChatRoleplay Q4_K_M·71.0 t/s	Q4_K_M	11.0 GB4%	71.0 t/s	131K	S86, Runs great
DeepSeek R1 Distill Qwen 1.5B1.5BChatReasoning Q4_K_M·579.1 t/s	Q4_K_M	2.6 GB1%	579.1 t/s	—	S100, Runs great
Qianfan OCR4.7BVision Q4_K_M·172.2 t/s	Q4_K_M	4.7 GB2%	172.2 t/s	33K	S98, Runs great
Sarashina2.2 0.5B Instruct v0.1793MChatFunctions Q4_K_M·637.0 t/s	Q4_K_M	1.3 GB1%	637.0 t/s	8K	S100, Runs great
Llm Jp 3.1 1.8B Instruct41.9BChat Q4_K_M·315.0 t/s	Q4_K_M	2.2 GB1%	315.0 t/s	4K	S100, Runs great
MiMo v2 Flash309.8BChatFunctions Q4_K_M·37.3 t/s	Q4_K_M	187.0 GB74%	37.3 t/s	262K	A74, Runs well
Scout 4B4.3BChatReasoningRoleplay Q4_K_M·176.9 t/s	Q4_K_M	3.6 GB1%	176.9 t/s	131K	S99, Runs great
Dream V0 Instruct 7B7.6BChatFunctions Q4_K_M·114.9 t/s	Q4_K_M	5.8 GB2%	114.9 t/s	131K	S94, Runs great
Gemma 3n E2B IT Litert Lm2BChat Q4_K_M·434.3 t/s	Q4_K_M	3.1 GB1%	434.3 t/s	—	S100, Runs great
UserLM 8B8.0BChat Q4_K_M·106.4 t/s	Q4_K_M	6.2 GB2%	106.4 t/s	8K	S92, Runs great

Mac Studio (M3 Ultra, 256GB) specs and power

Specs

Chip: M3 Ultra
CPU cores: 32
GPU cores: 80
Unified memory: 256 GB LPDDR5
For models: 252.5 GB (macOS keeps 3.5 GB)
Memory bandwidth: 819 GB/s
FP16 (dense): 65.5 TFLOPS
Neural Engine: 36 TOPS
NPU: 32-core Neural Engine
Power: 270 W max
Size: Mac Studio
Released: 2025-03-12
MSRP: $7,099

Mac Studio (M3 Ultra, 256GB) FAQ

Is the Mac Studio (M3 Ultra, 256GB) worth it for local LLMs in 2026?

At its $7,099 MSRP it gives 10.8 tok/s per $1,000 on Qwen3.6 35B A3B, 5th of the 6 machines compared here. It runs 27–31B and 120B MoE models well; dense 70B fits but is slow at 12 tok/s.

How much memory is available for AI on the Mac Studio (M3 Ultra, 256GB)?

About 252.5 GB of the 256 GB. macOS keeps roughly 3.5 GB for itself and other apps.

What is the largest model the Mac Studio (M3 Ultra, 256GB) can run?

MiniMax M3 (427.0B) is the largest at 4-bit or better: at Q4_K_S it uses 242.0 GB, 96% of the available memory, and runs at about 31.8 tok/s.

Can the Mac Studio (M3 Ultra, 256GB) run MiMo V2.6 Flash RL?

Yes. MiMo V2.6 Flash RL fits at Q4_K_M, using 186.5 GB of the 252.5 GB available, and it is the best model that runs well on this machine.

How long does the Mac Studio (M3 Ultra, 256GB) take to read a long prompt?

On MiMo V2.6 Flash RL, about 9.4 s for a 4K-token prompt and 75 s for a 32K-token document before the first word appears.

48 GB GPUs All Macs Desktops & mini PCs Best coding LLMs All hardware