How much memory does Gemma-4-26B-A4B at Q8_0 need?

28GB. That is 27GB of weights at 8.5 bits per weight, 0.3GB of K/V cache at 8K context, and 1.2GB of runtime overhead.

Which machine is cheapest for Gemma-4-26B-A4B at Q8_0?

Mac mini, M5 Pro, 48GB, $2,499 at Apple. It fits with 16GB of headroom and decodes at an est. 38.0 tok/s.

Cheapest box to run Gemma-4-26B-A4B at Q8_0

Google’s Gemma-4-26B-A4B is 25.2B parameters, 3.8B of them read per token. 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 32 minutes ago

Quant

Context

Gemma-4-26B-A4B at Q8_0 needs 28GB.

Cheapest Config that fits: Mac mini, M5 Pro, 48GB. Fits, 16GB headroom. Est. 38.0 tok/s.

Best offer today: $2,499 at Apple — in stock, checked 32 minutes ago.

Price history: 34 days recorded.

The 26B in the name is marketing; Google's own table reads 25.2B total and 3.8B active, and that is what we publish. Only the 5 global-attention layers cache unboundedly; the rest slide a 1024-token window.

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

Applelist pricein stockchecked 32 minutes ago$2,499

Mac mini, M6, 16GB$1,099
Won't fit−15GB headroom
Mac mini, M6, 24GB$1,299
Won't fit−7GB headroom
Mac mini, M6, 32GB$1,499
Tight1GB headroomest. 21.1 tok/s
Mac mini, M5 Pro, 24GB$1,699
Won't fit−7GB headroom
Mac mini, M5 Pro, 48GB$2,499
Fits16GB headroomest. 38.0 tok/s
Mac Studio, M5 Max, 36GB$2,499
Fits5GB headroomest. 57.0 tok/s
Mac mini, M5 Pro, 64GB$2,899
Fits31GB headroomest. 38.0 tok/s
GeForce RTX 5090, 32GB$2,999.99+host
Tight2GB headroomest. 222 tok/s
Mac Studio, M4 Max, 64GB, refurbished$3,049
Fits31GB headroomest. 67.6 tok/s
Mac Studio, M5 Max, 48GB$3,099
Fits16GB headroomest. 76.0 tok/s
Mac Studio, M5 Max, 64GB$3,499
Fits31GB headroomest. 76.0 tok/s
Ascent GX10, GB10 Grace Blackwell, 128GB$3,999.99
Fits92GB headroomest. 33.8 tok/s
Mac Studio, M3 Ultra, 96GB, refurbished$4,329
Fits60GB headroomest. 101 tok/s
Mac Studio, M4 Max, 128GB, refurbished$4,409
Fits92GB headroomest. 67.6 tok/s
DGX Spark, GB10 Grace Blackwell, 128GB$4,699
Fits92GB headroomest. 33.8 tok/s
Mac Studio, M5 Max, 128GB$5,099
Fits92GB headroomest. 76.0 tok/s
Mac Studio, M5 Ultra, 96GB$5,499
Fits60GB headroomest. 149 tok/s
Mac Studio, M3 Ultra, 256GB, refurbished$7,729
Fits220GB headroomest. 101 tok/s
Mac Studio, M5 Ultra, 256GB$9,499
Fits220GB headroomest. 149 tok/s
Mac Studio, M5 Ultra, 512GB—
Fits476GB headroomest. 149 tok/s

Config	Memory	Price	$/GB	Fit	Headroom	Est. speed
Mac mini, M6	16GB	$1,099	$69	Won't fit	−15GB	—
Mac mini, M6	24GB	$1,299	$54	Won't fit	−7GB	—
Mac mini, M6	32GB	$1,499	$47	Tight	1GB	21.1 tok/s
Mac mini, M5 Pro	24GB	$1,699	$71	Won't fit	−7GB	—
Mac mini, M5 Pro	48GB	$2,499	$52	Fits	16GB	38.0 tok/s
Mac Studio, M5 Max	36GB	$2,499	$69	Fits	5GB	57.0 tok/s
Mac mini, M5 Pro	64GB	$2,899	$45	Fits	31GB	38.0 tok/s
GeForce RTX 5090	32GB	$2,999.99+host	$94	Tight	2GB	222 tok/s
Mac Studio, M4 Max, refurb	64GB	$3,049	$48	Fits	31GB	67.6 tok/s
Mac Studio, M5 Max	48GB	$3,099	$65	Fits	16GB	76.0 tok/s
Mac Studio, M5 Max	64GB	$3,499	$55	Fits	31GB	76.0 tok/s
Ascent GX10, GB10 Grace Blackwell	128GB	$3,999.99	$31	Fits	92GB	33.8 tok/s
Mac Studio, M3 Ultra, refurb	96GB	$4,329	$45	Fits	60GB	101 tok/s
Mac Studio, M4 Max, refurb	128GB	$4,409	$34	Fits	92GB	67.6 tok/s
DGX Spark, GB10 Grace Blackwell	128GB	$4,699	$37	Fits	92GB	33.8 tok/s
Mac Studio, M5 Max	128GB	$5,099	$40	Fits	92GB	76.0 tok/s
Mac Studio, M5 Ultra	96GB	$5,499	$57	Fits	60GB	149 tok/s
Mac Studio, M3 Ultra, refurb	256GB	$7,729	$30	Fits	220GB	101 tok/s
Mac Studio, M5 Ultra	256GB	$9,499	$37	Fits	220GB	149 tok/s
Mac Studio, M5 Ultra	512GB	—	—	Fits	476GB	149 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 = 25.2B × 1.063 B/param (8.5 bits) + 0.042 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (3.8B active × 1.063 B/param) = 4.0 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

Qwen3-Coder-NextQ3_K_M, est. 105 tok/s
Llama-3.3-70BQ3_K_M, est. 5.8 tok/s
Qwen3.5-35B-A3BQ8_0, est. 48.2 tok/s
Qwen3-32BQ8_0, est. 5.9 tok/s
Gemma-4-31BQ8_0, est. 6.1 tok/s
Qwen3-30B-A3BQ8_0, est. 43.8 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.