Local AI hardware, model first
Memory needs for Llama, Qwen, Gemma, gpt-oss, DeepSeek, FLUX and Whisper, matched to new-retail PCs, mini PCs and parts. Every number is cited, and every price shows when we checked it.
We earn from qualifying purchases via affiliate links. Disclosure.
What do you want to run?
Pick a model to see the smallest tier that runs it.
Mini PCs · 128 GB unified
$3,919 $4,899 save $980
Was = Minisforum list price
Runs: gpt-oss 120B, Qwen3.5 122B-A10B, Llama 3.3 70B. Unified 128 GB tier
Store lists estimated shipping as late November.
Minisforum · checked
See deal at Minisforum (opens in a new tab)
GPUs · 12 GB VRAM
$329.99 $370.99 save $41
Was = Newegg strike-through price
Runs: Llama 3.1 8B, Phi-4, Qwen3 14B, SDXL, Whisper. 8–12 GB GPU tier
Sold by Newegg. Arc support varies between AI tools.
Newegg · checked
See deal at Newegg (opens in a new tab)
GPUs · 16 GB VRAM
$799.99 $847.99 save $48
Was = Newegg strike-through price
Runs: gpt-oss 20B, Mistral Small 3.2 24B (tight), FLUX at FP8. 16 GB GPU tier
Small discount on a high price; compare other 16 GB cards first.
Newegg · checked
See deal at Newegg (opens in a new tab)
Mini PCs · 16 GB system RAM
$615 $769 save $154
Was = Minisforum list price
Runs: Llama 3.1 8B and Whisper turbo on the 780M iGPU. Mini PC guide
Minisforum · checked
See deal at Minisforum (opens in a new tab)
Networking · Networking
$59.99 $129.99 save $70
Was = Newegg strike-through price
Runs: Links 2.5GbE mini PCs and NAS boxes; helps with model copies and clustering. Clustering guide
The listing's seller was not confirmed.
Newegg · checked
See deal at Newegg (opens in a new tab)
Mini PCs · 64 GB unified
$2,399 $2,999 save $600
Was = Minisforum list price
Runs: Llama 3.3 70B (tight), Qwen3.5 35B-A3B, Gemma 4 31B. Unified 64 GB tier
A NAS chassis: drives are extra.
Minisforum · checked
See deal at Minisforum (opens in a new tab)
8–12 GB 8B–14B chat models at Q4, SDXL, Whisper, FLUX at Q4
From $329.99ASRock Challenger Intel Arc B580 12GB (card only) · checked Oct 27, 2026, about 12:35 PM CT
16 GB gpt-oss 20B, Mistral Small 3.2 24B (tight) and FLUX at FP8, plus everything in the 8–12 GB tier
From $1,599.99ABS Cyclone Aqua gaming PC (Core i7-14700F, RTX 5060 Ti 16GB, 16GB DDR5, 1TB SSD) · checked Oct 27, 2026, about 12:35 PM CT
27B–35B models at Q4 (Qwen3.5 27B, Gemma 4 31B, DeepSeek-R1 32B)
Check price at store
Unified 64 35B MoE models; Llama 3.3 70B at Q4 (tight)
From $2,599Minisforum MS-S1 MAX 64GB / 2TB (Ryzen AI Max+ 395) · checked Oct 27, 2026, 12:40 PM CT
Unified 128 gpt-oss 120B, Qwen3.5 122B-A10B, Llama 3.3 70B with room
From $3,919Minisforum MS-S1 MAX 128GB / 2TB (Ryzen AI Max+ 395) · checked Oct 27, 2026, 12:40 PM CT
gpt-oss 20B is a 14 GB download in Ollama. See the 16 GB and 24 GB GPUs, prebuilts and parts that fit it, with prices checked Oct 27, 2026.
Llama 3.3 70B is a 43 GB download at Q4_K_M. Compare 64 GB and 128 GB unified-memory mini PCs with a two-GPU build.
Qwen3.5 27B needs 17–20 GB, 35B-A3B needs 22 GB and 122B-A10B needs 81 GB in Ollama. Here is the hardware tier for each size.
Gemma 4 31B is a 19–20 GB download in Ollama with a 256K context window. See the GPUs and systems that hold it.
The DeepSeek-R1 32B distill is a 20 GB download in Ollama. See the 24–32 GB GPUs and unified-memory boxes that run it.
SDXL runs on 8 GB of VRAM per Stability AI; FLUX.1 dev needs about 12 GB at FP8 (third-party estimate). Pick the GPU tier for image generation.
