How much memory does Llama-3.3-70B at Q4_K_M need?

46GB. That is 42GB of weights at 4.83 bits per weight, 2.6GB of K/V cache at 8K context, and 1.2GB of runtime overhead.

Which machine is cheapest for Llama-3.3-70B at Q4_K_M?

Mac mini, M5 Pro, 64GB, $2,899 at Apple. It fits with 13GB of headroom and decodes at an est. 4.7 tok/s.

Cheapest box to run Llama-3.3-70B at Q4_K_M

Meta’s Llama-3.3-70B is 70B parameters, dense. The number below is computed from that and each vendor’s published specs, not asserted. Switch the quant or the context and it recomputes in place.

Fit finderprices checked 9 hours ago

Quant

Context

Llama-3.3-70B at Q4_K_M needs 46GB.

Cheapest Config that fits: Mac mini, M5 Pro, 64GB. Fits, 13GB headroom. Est. 4.7 tok/s.

Best offer today: $2,899 at Apple — in stock, checked 9 hours ago.

Collecting price history — 1 of 90 days. No timing verdict yet.

Offers — Mac mini, M5 Pro, 64GB, 512GB

Applelist pricein stockchecked 9 hours ago$2,899

Mac mini, M6, 16GB$1,099
Won't fit−33GB headroom
Mac mini, M6, 24GB$1,299
Won't fit−25GB headroom
Mac mini, M6, 32GB$1,499
Won't fit−17GB headroom
Mac mini, M5 Pro, 24GB$1,699
Won't fit−25GB headroom
Mac mini, M5 Pro, 48GB$2,499
Won't fit−2GB headroom
Mac Studio, M5 Max, 36GB$2,499
Won't fit−13GB headroom
Mac mini, M5 Pro, 64GB$2,899
Fits13GB headroomest. 4.7 tok/s
Mac Studio, M5 Max, 48GB$3,099
Won't fit−2GB headroom
Mac Studio, M5 Max, 64GB$3,499
Fits13GB headroomest. 9.4 tok/s
DGX Spark, GB10 Grace Blackwell, 128GB$4,699
Fits74GB headroomest. 4.2 tok/s
Mac Studio, M5 Max, 128GB$5,099
Fits74GB headroomest. 9.4 tok/s
Mac Studio, M5 Ultra, 96GB$5,499
Fits42GB headroomest. 18.5 tok/s
GeForce RTX 5090, 32GB$5,699.99+host
Won't fit−15GB headroom
Mac Studio, M5 Ultra, 256GB$9,499
Fits202GB headroomest. 18.5 tok/s
Mac Studio, M5 Ultra, 512GB—
Fits458GB headroomest. 18.5 tok/s

Config	Memory	Price	$/GB	Fit	Headroom	Est. speed
Mac mini, M6	16GB	$1,099	$69	Won't fit	−33GB	—
Mac mini, M6	24GB	$1,299	$54	Won't fit	−25GB	—
Mac mini, M6	32GB	$1,499	$47	Won't fit	−17GB	—
Mac mini, M5 Pro	24GB	$1,699	$71	Won't fit	−25GB	—
Mac mini, M5 Pro	48GB	$2,499	$52	Won't fit	−2GB	—
Mac Studio, M5 Max	36GB	$2,499	$69	Won't fit	−13GB	—
Mac mini, M5 Pro	64GB	$2,899	$45	Fits	13GB	4.7 tok/s
Mac Studio, M5 Max	48GB	$3,099	$65	Won't fit	−2GB	—
Mac Studio, M5 Max	64GB	$3,499	$55	Fits	13GB	9.4 tok/s
DGX Spark, GB10 Grace Blackwell	128GB	$4,699	$37	Fits	74GB	4.2 tok/s
Mac Studio, M5 Max	128GB	$5,099	$40	Fits	74GB	9.4 tok/s
Mac Studio, M5 Ultra	96GB	$5,499	$57	Fits	42GB	18.5 tok/s
GeForce RTX 5090	32GB	$5,699.99+host	$178	Won't fit	−15GB	—
Mac Studio, M5 Ultra	256GB	$9,499	$37	Fits	202GB	18.5 tok/s
Mac Studio, M5 Ultra	512GB	—	—	Fits	458GB	18.5 tok/s

GeForce RTX 5090 is a card, not a computer: its price is the US street median plus $1,000 for the minimum host machine. NVIDIA's $1,999 MSRP is not a price anyone pays.

memory = 70B × 0.604 B/param (4.83 bits) + 0.328 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.65 ÷ (70B active × 0.604 B/param) = 42.3 GB read per token
usable = memory − operating system reserve (8% on a Mac, 6% on a DGX Spark, 4% on a card)

Bandwidth is the vendor’s published figure. Every speed is an estimate and says so. Methodology.

What else Mac mini M5 Pro 64GB runs

Llama-4-Scout-17B-16EQ3_K_M, est. 18.5 tok/s
Llama-3.3-70BQ5_K_M, est. 4.0 tok/s
Qwen3-32BQ8_0, est. 5.9 tok/s
Gemma-3-27BBF16, est. 3.7 tok/s
gpt-oss-20bBF16, est. 21.3 tok/s

Best quant that still fits at 8K context, largest model first. Full list on the Config page.

Bandwidth 307GB/s, usable memory 59GB of 64GB.