How much memory does gpt-oss-20b at Q3_K_M need?

12GB. That is 10GB of weights at 3.91 bits per weight, 0.4GB of K/V cache at 8K context, and 1.2GB of runtime overhead.

Which machine is cheapest for gpt-oss-20b at Q3_K_M?

Mac mini, M6, 24GB, $1,299 at Expercom. It fits with 9GB of headroom and decodes at an est. 48.3 tok/s.

Cheapest box to run gpt-oss-20b at Q3_K_M

OpenAI’s gpt-oss-20b is 21B parameters, 3.6B 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 18 days ago

Quant

Context

gpt-oss-20b at Q3_K_M needs 12GB.

Cheapest Config that fits: Mac mini, M6, 24GB. Fits, 9GB headroom. Est. 48.3 tok/s.

Best offer today: $1,299 at Expercom — in stock, checked 18 days ago.

Price history: 47 days recorded.

Offers — Mac mini, M6, 24GB, 512GB

ExpercomApple's own price, to the cent.in stockchecked 18 days ago$1,299
Applelist pricein stockchecked 1 hour ago$1,299
Best BuyApple's own price, to the cent.in stockchecked 46 days ago$1,299

Mac mini, M6, 16GB$1,099
Tight1GB headroomest. 48.3 tok/s
Mac mini, M6, 24GB$1,299
Fits9GB headroomest. 48.3 tok/s
Mac mini, M6, 32GB$1,499
Fits17GB headroomest. 48.3 tok/s
Mac mini, M5 Pro, 24GB$1,699
Fits9GB headroomest. 87.2 tok/s
Mac mini, M5 Pro, 48GB$2,499
Fits32GB headroomest. 87.2 tok/s
Mac Studio, M5 Max, 36GB$2,499
Fits21GB headroomest. 131 tok/s
Mac mini, M5 Pro, 64GB$2,899
Fits47GB headroomest. 87.2 tok/s
GeForce RTX 5090, 32GB$2,999.99+host
Fits19GB headroomest. 509 tok/s
Mac Studio, M4 Max, 64GB, refurbished$3,049
Fits47GB headroomest. 155 tok/s
Mac Studio, M5 Max, 48GB$3,099
Fits32GB headroomest. 174 tok/s
Mac Studio, M5 Max, 64GB$3,499
Fits47GB headroomest. 174 tok/s
Ascent GX10, GB10 Grace Blackwell, 128GB$3,999.99
Fits108GB headroomest. 77.6 tok/s
Mac Studio, M3 Ultra, 96GB, refurbished$4,329
Fits76GB headroomest. 233 tok/s
Mac Studio, M4 Max, 128GB, refurbished$4,409
Fits108GB headroomest. 155 tok/s
DGX Spark, GB10 Grace Blackwell, 128GB$4,699
Fits108GB headroomest. 77.6 tok/s
Mac Studio, M5 Max, 128GB$5,099
Fits108GB headroomest. 174 tok/s
Mac Studio, M5 Ultra, 96GB$5,499
Fits76GB headroomest. 341 tok/s
Mac Studio, M3 Ultra, 256GB, refurbished$7,729
Fits236GB headroomest. 233 tok/s
Mac Studio, M5 Ultra, 256GB$9,499
Fits236GB headroomest. 341 tok/s
Mac Studio, M5 Ultra, 512GB—
Fits492GB headroomest. 341 tok/s

Config	Memory	Price	Fit	Headroom	Est. speed
Mac mini, M6	16GB	$1,099	Tight	1GB	48.3 tok/s
Mac mini, M6	24GB	$1,299	Fits	9GB	48.3 tok/s
Mac mini, M6	32GB	$1,499	Fits	17GB	48.3 tok/s
Mac mini, M5 Pro	24GB	$1,699	Fits	9GB	87.2 tok/s
Mac mini, M5 Pro	48GB	$2,499	Fits	32GB	87.2 tok/s
Mac Studio, M5 Max	36GB	$2,499	Fits	21GB	131 tok/s
Mac mini, M5 Pro	64GB	$2,899	Fits	47GB	87.2 tok/s
GeForce RTX 5090	32GB	$2,999.99+host	Fits	19GB	509 tok/s
Mac Studio, M4 Max, refurb	64GB	$3,049	Fits	47GB	155 tok/s
Mac Studio, M5 Max	48GB	$3,099	Fits	32GB	174 tok/s
Mac Studio, M5 Max	64GB	$3,499	Fits	47GB	174 tok/s
Ascent GX10, GB10 Grace Blackwell	128GB	$3,999.99	Fits	108GB	77.6 tok/s
Mac Studio, M3 Ultra, refurb	96GB	$4,329	Fits	76GB	233 tok/s
Mac Studio, M4 Max, refurb	128GB	$4,409	Fits	108GB	155 tok/s
DGX Spark, GB10 Grace Blackwell	128GB	$4,699	Fits	108GB	77.6 tok/s
Mac Studio, M5 Max	128GB	$5,099	Fits	108GB	174 tok/s
Mac Studio, M5 Ultra	96GB	$5,499	Fits	76GB	341 tok/s
Mac Studio, M3 Ultra, refurb	256GB	$7,729	Fits	236GB	233 tok/s
Mac Studio, M5 Ultra	256GB	$9,499	Fits	236GB	341 tok/s
Mac Studio, M5 Ultra	512GB	—	Fits	492GB	341 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 = 21B × 0.489 B/param (3.91 bits) + 0.049 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (3.6B active × 0.489 B/param) = 1.8 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 M6 24GB runs

Qwen3.5-35B-A3BQ3_K_M, est. 58.0 tok/s
Qwen3-32BQ3_K_M, est. 7.1 tok/s
Gemma-4-31BQ3_K_M, est. 7.4 tok/s
Qwen3-30B-A3BQ4_K_M, est. 42.7 tok/s
Gemma-3-27BQ4_K_M, est. 6.8 tok/s
Qwen3.6-27BQ5_K_M, est. 5.8 tok/s

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

Bandwidth 170GB/s, usable memory 21GB of 24GB.