TokenAssemble

HP · unified memory

HP Z2 Mini G1a Ryzen AI Max+ PRO 395 128GB

The HP Z2 Mini G1a Ryzen AI Max+ PRO 395 128GB has 128GB of unified memory, roughly 96GB of it GPU-addressable, with 256 GB/s of bandwidth. The largest model that loads cleanly at a quality quant is gpt-oss-120b (llama.cpp, 8K context).

Unified memory: 128GB shared between CPU and GPU; we count ~75% as GPU-addressable for model weights. This class fits huge models but is bandwidth-bound: a box that fits a 120B-class model can still decode slowly, while prompt processing may land a higher tier — one headline number would mislead in both directions, so speed here is a tier, labeled beta until calibrated.
Unified memory
128 GB
GPU-addressable
~96 GB (75%)
Memory bandwidth
256 GB/s
Backend
rocm
Price (as configured)
$5,021

Specs as of 2026-07-11 · source

Affiliate link — we may earn a commission. Verdicts are computed before any link is attached.

What an HP Z2 Mini G1a Ryzen AI Max+ PRO 395 128GB can run

Every model we track, graded at its best clean-fit quant.

ModelBest-quant verdict · llama.cpp · 8K context
Qwen3-8BFits · A+ · Q8_0usablebeta
gemma-4-26B-A4B-itFits · A+ · Q8_0usablebeta
Qwen2.5-7B-InstructFits · A+ · Q8_0usablebeta
Qwen2.5-1.5B-InstructFits · A+ · Q8_0interactivebeta
Llama-3.2-1B-InstructFits · A+ · Q8_0interactivebeta
Llama-3.1-8B-InstructFits · A+ · Q8_0usablebeta
DeepSeek-R1Won't fitbeta
Qwen3.5-9BFits · A+ · Q8_0usablebeta
gpt-oss-20bFits · A+ · MXFP4interactivebeta
Qwen3.6-35B-A3BFits · A+ · Q8_0interactivebeta
Qwen2.5-3B-InstructFits · A+ · Q8_0interactivebeta
Qwen3-32BFits · A+ · Q8_0painfulbeta
Qwen3.6-27BFits · A+ · Q8_0painfulbeta
Qwen2.5-0.5B-InstructFits · A+ · Q8_0interactivebeta
Mistral-7B-Instruct-v0.3Fits · A+ · Q8_0usablebeta
gpt-oss-120bFits · A+ · MXFP4interactivebeta
Qwen3-14BFits · A+ · Q8_0usablebeta
Qwen3-30B-A3BFits · A+ · Q8_0interactivebeta
Qwen2.5-32B-InstructFits · A+ · Q8_0painfulbeta
DeepSeek-V4-FlashWon't fitbeta
Qwen2.5-14B-InstructFits · A+ · Q8_0usablebeta
Llama-3.2-3B-InstructFits · A+ · Q8_0interactivebeta
gemma-3-4b-itFits · A+ · Q8_0interactivebeta
gemma-3-12b-itFits · A+ · Q8_0usablebeta
Llama-3.1-70B-InstructFits · B · Q8_0painfulbeta
Phi-3.5-mini-instructFits · A+ · Q8_0interactivebeta
Phi-4Fits · A+ · Q8_0usablebeta
DeepSeek-R1-Distill-Qwen-32BFits · A+ · Q8_0painfulbeta
gemma-3-27b-itFits · A+ · Q8_0painfulbeta
Llama-3.3-70B-InstructFits · B · Q8_0painfulbeta
Qwen2.5-72B-InstructFits · B · Q8_0painfulbeta
gemma-2-2b-itFits · A+ · Q8_0interactivebeta
DeepSeek-R1-Distill-Qwen-14BFits · A+ · Q8_0usablebeta
DeepSeek-R1-Distill-Llama-8BFits · A+ · Q8_0usablebeta
DeepSeek-R1-Distill-Qwen-7BFits · A+ · Q8_0usablebeta
gemma-2-9b-itFits · A+ · Q8_0usablebeta
Mistral-Nemo-Instruct-2407Fits · A+ · Q8_0usablebeta
QwQ-32BFits · A+ · Q8_0painfulbeta
Mistral-Small-24B-Instruct-2501Fits · A+ · Q8_0painfulbeta
gemma-2-27b-itFits · A+ · Q8_0painfulbeta

Verdicts computed from published specs and measured GGUF sizes — how the math works.