Local LLM configurator

Model

Quantization

Q4_K_M is the right trade-off for most uses. What is this?

Target speed

Around 10 tok/s reads comfortably; 30+ feels instant.

Platform

Advanced options8K tokens · KV cache f16

The defaults suit most uses. You can skip this section.

Context length

How much text the model keeps in memory. 8K is enough for chat; long documents need 32K or more. What is this?

KV cache type

Leave f16 if unsure. q8_0 roughly halves the conversation's memory, q4_0 nearly quarters it; the option must also be enabled in your app. What is this?

Compatible machines

What the model needs

Qwen3.5 122B-A10B Q4_K_M with a 8K context needs about 76.3 GB (up to 81.2 GB) of GPU or unified memory.

Compatible configurations for Qwen3.5 122B-A10B Q4_K_M · 8K context
	NVIDIARTX PRO 6000 BlackwellExcellent RTX PRO 6000 Blackwell Workstation Edition · 96 GB GDDR7 ECC 1,792 GB/s System RAM 64 GB Smallest configuration of this platform that runs the model entirely in fast memory.	AMD2 × Radeon PRO W7900Excellent Radeon PRO W7900 · 48 GB GDDR6 ECC 864 GB/s System RAM 64 GB Smallest configuration of this platform that runs the model entirely in fast memory.	Apple SiliconM1 Ultra · 128 GBExcellent Mac Studio 64 GPU · 800 GB/s Smallest configuration of this platform that runs the model entirely in fast memory.	Unified memory PCNVIDIA DGX Spark (GB10) · 128 GBExcellent GB10 Grace Blackwell 273 GB/s Smallest configuration of this platform that runs the model entirely in fast memory.	Lowest memory tierA100 80GBCompatible with limits A100 80GB PCIe · 80 GB HBM2e 1,935 GB/s System RAM 64 GB Least memory among all configurations that fit.	Highest estimated speedH100 PCIeCompatible with limits H100 PCIe 80GB · 80 GB HBM2e 2,000 GB/s System RAM 64 GB Highest estimated generation speed among configurations that fit.
Memory for the model	93 GB	93 GB	96 GB	109 GBMost memory	78 GB	78 GB
Estimated speed	102–164 tok/sestimate	38–73 tok/sestimate	40–66 tok/sestimate	25–34 tok/sestimate	107–172 tok/sestimate	109–176 tok/sestimateFastest
Power draw (GPU/board)	~600 W	~590 W	–	~240 WLowest power	~300 W	~350 W
Price	No current offer	No current offer	No current offer	No current offer	No current offer	No current offer
Confidence	low	low	low	low	low	low

Leaders are marked per criterion; there is no overall winner. Speeds are estimates unless marked measured. Power is the GPU's rated board power (or the platform's, when known).

Suggested PC build

GPU: RTX PRO 6000 Blackwell Workstation Edition
Processor: Ryzen 7 9700X
RAM: 96 GB DDR5
Power supply: 1000 W

The processor matters little when the model runs on the GPU; any current 8-core CPU is enough. PSU sized with 30% headroom for power spikes.

Compatibility checks: No compatibility problem detected between these parts.

Where to buy

Your market

Prices and links follow the market you choose.

RTX PRO 6000 Blackwell Workstation Edition

RTX PRO 6000 Blackwell Workstation Edition · 96 GB GDDR7 ECC · 1,792 GB/s

ExcellentRuns Qwen3.5 122B-A10B Q4_K_M entirely in memory

Price not available

Availability unknown

View on amazon.com

Affiliate link

Radeon PRO W7900

Radeon PRO W7900 · 48 GB GDDR6 ECC · 864 GB/s

ExcellentRuns Qwen3.5 122B-A10B Q4_K_M entirely in memory

Price not available

Availability unknown

View on amazon.com

Affiliate link

NVIDIA DGX Spark (GB10)

GB10 Grace Blackwell · 273 GB/s

ExcellentRuns Qwen3.5 122B-A10B Q4_K_M entirely in memory

Price not available

Availability unknown

View on amazon.com

Affiliate link

No current offer imported for this market: the link opens a marketplace search.