How much memory does Qwen3-32B at Q5_K_M need?

26GB. That is 23GB of weights at 5.67 bits per weight, 2.1GB of K/V cache at 8K context, and 1.2GB of runtime overhead.

Which machine is cheapest for Qwen3-32B at Q5_K_M?

Mac mini, M5 Pro, 48GB, $2,499 at Apple. It fits with 18GB of headroom and decodes at an est. 8.8 tok/s.

Cheapest box to run Qwen3-32B at Q5_K_M

Alibaba’s Qwen3-32B is 32B 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

Qwen3-32B at Q5_K_M needs 26GB.

Cheapest Config that fits: Mac mini, M5 Pro, 48GB. Fits, 18GB headroom. Est. 8.8 tok/s.

Best offer today: $2,499 at Apple — in stock, checked 9 hours ago.

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

Offers — Mac mini, M5 Pro, 48GB, 512GB

Applelist pricein stockchecked 9 hours ago$2,499

Mac mini, M6, 16GB$1,099
Won't fit−13GB headroom
Mac mini, M6, 24GB$1,299
Won't fit−5GB headroom
Mac mini, M6, 32GB$1,499
Tight3GB headroomest. 4.9 tok/s
Mac mini, M5 Pro, 24GB$1,699
Won't fit−5GB headroom
Mac mini, M5 Pro, 48GB$2,499
Fits18GB headroomest. 8.8 tok/s
Mac Studio, M5 Max, 36GB$2,499
Fits7GB headroomest. 17.6 tok/s
Mac mini, M5 Pro, 64GB$2,899
Fits33GB headroomest. 8.8 tok/s
Mac Studio, M5 Max, 48GB$3,099
Fits18GB headroomest. 17.6 tok/s
Mac Studio, M5 Max, 64GB$3,499
Fits33GB headroomest. 17.6 tok/s
DGX Spark, GB10 Grace Blackwell, 128GB$4,699
Fits94GB headroomest. 7.8 tok/s
Mac Studio, M5 Max, 128GB$5,099
Fits94GB headroomest. 17.6 tok/s
Mac Studio, M5 Ultra, 96GB$5,499
Fits62GB headroomest. 34.4 tok/s
GeForce RTX 5090, 32GB$5,699.99+host
Fits5GB headroomest. 51.4 tok/s
Mac Studio, M5 Ultra, 256GB$9,499
Fits222GB headroomest. 34.4 tok/s
Mac Studio, M5 Ultra, 512GB—
Fits478GB headroomest. 34.4 tok/s

Config	Memory	Price	$/GB	Fit	Headroom	Est. speed
Mac mini, M6	16GB	$1,099	$69	Won't fit	−13GB	—
Mac mini, M6	24GB	$1,299	$54	Won't fit	−5GB	—
Mac mini, M6	32GB	$1,499	$47	Tight	3GB	4.9 tok/s
Mac mini, M5 Pro	24GB	$1,699	$71	Won't fit	−5GB	—
Mac mini, M5 Pro	48GB	$2,499	$52	Fits	18GB	8.8 tok/s
Mac Studio, M5 Max	36GB	$2,499	$69	Fits	7GB	17.6 tok/s
Mac mini, M5 Pro	64GB	$2,899	$45	Fits	33GB	8.8 tok/s
Mac Studio, M5 Max	48GB	$3,099	$65	Fits	18GB	17.6 tok/s
Mac Studio, M5 Max	64GB	$3,499	$55	Fits	33GB	17.6 tok/s
DGX Spark, GB10 Grace Blackwell	128GB	$4,699	$37	Fits	94GB	7.8 tok/s
Mac Studio, M5 Max	128GB	$5,099	$40	Fits	94GB	17.6 tok/s
Mac Studio, M5 Ultra	96GB	$5,499	$57	Fits	62GB	34.4 tok/s
GeForce RTX 5090	32GB	$5,699.99+host	$178	Fits	5GB	51.4 tok/s
Mac Studio, M5 Ultra	256GB	$9,499	$37	Fits	222GB	34.4 tok/s
Mac Studio, M5 Ultra	512GB	—	—	Fits	478GB	34.4 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 = 32B × 0.709 B/param (5.67 bits) + 0.262 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.65 ÷ (32B active × 0.709 B/param) = 22.7 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 48GB runs

Llama-3.3-70BQ3_K_M, est. 5.8 tok/s
Qwen3-32BQ8_0, est. 5.9 tok/s
Gemma-3-27BQ8_0, est. 7.0 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 44GB of 48GB.