Best home rig for Qwen3.5 122B

Will it fit my GPU? Best value GPU per euro →

Real, measured generation speeds for Qwen3.5 122B on home hardware, fastest first. Different quants/engines aren't directly comparable.

Build	Quant	Engine	tok/s	Cost
Microsoft Surface Laptop Ultra (RTX Spark N1X): llama.cpp speeds on 2026 models (Micro Center)	A10B, general, short prompt	llama.cpp (Nvidia ARM64 build)	35.2		source

Rent before you buy: try this model on a cloud GPU for a few dollars first: Runpod (referral link; sign up with Google to get the new-user bonus)

Memory math: how much VRAM / RAM does Qwen3.5 122B need?

Qwen3.5 122B has 122B parameters (source: Ollama library tag 122b (nominal)). Weight size = parameters × bits-per-weight ÷ 8, before context (KV cache) and runtime overhead.

Format	Bits/weight	Weights	Smallest common memory size that holds the weights alone
Q8_0	8.50	129.6 GB	192 GB
Q6_K	6.56	100.1 GB	128 GB
Q5_K_M	5.70	87.0 GB	96 GB
Q4_K_M	4.89	74.6 GB	96 GB
Q3_K_M	4.00	60.9 GB	64 GB
Q2_K	3.16	48.2 GB	64 GB

Bits-per-weight for GGUF quants are llama.cpp's own figures (llama.cpp quantize README). "As measured" rows apply the source's stated bits-per-weight to the parameter count; real files differ slightly because some layers stay at higher precision. Try other combinations in the VRAM calculator.

FAQ

How much VRAM does Qwen3.5 122B need?

The weights are about 74.6 GB at Q4_K_M (122B params × 4.89 bits ÷ 8, before context/KV cache). Add memory for context (KV cache) and the runtime.

What is the fastest measured hardware for Qwen3.5 122B?

RTX Spark N1X: 35.2 tok/s generation (llama.cpp (Nvidia ARM64 build), A10B, general, short prompt), from https://www.microcenter.com/site/mc-news/article/benchmarking-the-rtx-spark.aspx.

Which GPUs have measured results for Qwen3.5 122B?

RTX Spark N1X.

Can it run on… RTX Spark N1X