How much memory does Gemma-3-27B at Q6_K need?

25GB. That is 22GB of weights at 6.56 bits per weight, 2.0GB of K/V cache at 8K context, and 1.2GB of runtime overhead.

Which machine is cheapest for Gemma-3-27B at Q6_K?

Mac mini, M5 Pro, 48GB, $2,499 at Apple. It fits with 19GB of headroom and decodes at an est. 9.0 tok/s.

Cheapest box to run Gemma-3-27B at Q6_K

Google’s Gemma-3-27B is 27B 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 10 hours ago

Quant

Context

Gemma-3-27B at Q6_K needs 25GB.

Cheapest Config that fits: Mac mini, M5 Pro, 48GB. Fits, 19GB headroom. Est. 9.0 tok/s.

Best offer today: $2,499 at Apple — in stock, checked 10 hours ago.

Collecting price history — 2 of 90 days. No timing verdict yet.

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

Applelist pricein stockchecked 10 hours ago$2,499

Mac mini, M6, 16GB$1,099
Won't fit−12GB headroom
Mac mini, M6, 24GB$1,299
Won't fit−4GB headroom
Mac mini, M6, 32GB$1,499
Tight4GB headroomest. 5.0 tok/s
Mac mini, M5 Pro, 24GB$1,699
Won't fit−4GB headroom
Mac mini, M5 Pro, 48GB$2,499
Fits19GB headroomest. 9.0 tok/s
Mac Studio, M5 Max, 36GB$2,499
Fits8GB headroomest. 18.0 tok/s
Mac mini, M5 Pro, 64GB$2,899
Fits34GB headroomest. 9.0 tok/s
GeForce RTX 5090, 32GB$2,999.99+host
Fits5GB headroomest. 52.6 tok/s
Mac Studio, M5 Max, 48GB$3,099
Fits19GB headroomest. 18.0 tok/s
Mac Studio, M5 Max, 64GB$3,499
Fits34GB headroomest. 18.0 tok/s
Ascent GX10, GB10 Grace Blackwell, 128GB$3,999.99
Fits95GB headroomest. 8.0 tok/s
DGX Spark, GB10 Grace Blackwell, 128GB$4,699
Fits95GB headroomest. 8.0 tok/s
Mac Studio, M5 Max, 128GB$5,099
Fits95GB headroomest. 18.0 tok/s
Mac Studio, M5 Ultra, 96GB$5,499
Fits63GB headroomest. 35.2 tok/s
Mac Studio, M5 Ultra, 256GB$9,499
Fits223GB headroomest. 35.2 tok/s
Mac Studio, M5 Ultra, 512GB—
Fits479GB headroomest. 35.2 tok/s

Config	Memory	Price	$/GB	Fit	Headroom	Est. speed
Mac mini, M6	16GB	$1,099	$69	Won't fit	−12GB	—
Mac mini, M6	24GB	$1,299	$54	Won't fit	−4GB	—
Mac mini, M6	32GB	$1,499	$47	Tight	4GB	5.0 tok/s
Mac mini, M5 Pro	24GB	$1,699	$71	Won't fit	−4GB	—
Mac mini, M5 Pro	48GB	$2,499	$52	Fits	19GB	9.0 tok/s
Mac Studio, M5 Max	36GB	$2,499	$69	Fits	8GB	18.0 tok/s
Mac mini, M5 Pro	64GB	$2,899	$45	Fits	34GB	9.0 tok/s
GeForce RTX 5090	32GB	$2,999.99+host	$94	Fits	5GB	52.6 tok/s
Mac Studio, M5 Max	48GB	$3,099	$65	Fits	19GB	18.0 tok/s
Mac Studio, M5 Max	64GB	$3,499	$55	Fits	34GB	18.0 tok/s
Ascent GX10, GB10 Grace Blackwell	128GB	$3,999.99	$31	Fits	95GB	8.0 tok/s
DGX Spark, GB10 Grace Blackwell	128GB	$4,699	$37	Fits	95GB	8.0 tok/s
Mac Studio, M5 Max	128GB	$5,099	$40	Fits	95GB	18.0 tok/s
Mac Studio, M5 Ultra	96GB	$5,499	$57	Fits	63GB	35.2 tok/s
Mac Studio, M5 Ultra	256GB	$9,499	$37	Fits	223GB	35.2 tok/s
Mac Studio, M5 Ultra	512GB	—	—	Fits	479GB	35.2 tok/s

GeForce RTX 5090 is a card, not a computer: every price shown for it is the cheapest listing we found plus $1,000 for the minimum host machine.

memory = 27B × 0.820 B/param (6.56 bits) + 0.254 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.65 ÷ (27B active × 0.820 B/param) = 22.1 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.