Home›Hardware›Mini-PC with an NPU and 32 GB LPDDR5X
Mini-PC with an NPU and 32 GB LPDDR5X
mini-PC with NPU32 GiB120 GB/s4.0 TFLOPS
This machine leaves 29.00 GiB for the model once the system has its share, and it takes 14 of the 18 models we measure at the reference compression. What decides how fast it answers is not the chip: it is the 120 GB/s of memory bandwidth.
Left for the model
Estimated
Memory bandwidth
Estimated
Models that fit
Measured by Local AI Scope
The memory you can actually use
Unified memory is shared with the operating system and everything you have open. The reserve here is deliberately conservative; you can change it in the tool.
| Memory | GiB |
|---|---|
| Memory on the box | 32 |
| Reserved for the system | −3.00 |
| Left for the model | 29.00 |
Memory bandwidth
Two different limits, and people confuse them constantly. While the model is writing its answer, bandwidth rules: every single token means reading the weights again, so tokens per second is roughly bandwidth divided by model size. While the model is reading your document, compute rules: the whole input is processed at once. That is why a laptop with no GPU can take minutes before the first word appears and then type at a tolerable pace.
120 GB/s — Estimated · no single manufacturer: this is a class of machine, not a product. The figure is a conservative estimate for that class, not a specification.
Compute: 4.0 TFLOPS (estimate for this class of machine) — Estimated · no single manufacturer: this is a class of machine, not a product. The figure is a conservative estimate for that class, not a specification. The figures in this line do not all come from the same measure across machines, because each manufacturer publishes a different one. Compare them between machines only with that in mind.
Which models fit here, and which do not
Fitting means three things at once fit in the memory left over: the model file, its context memory, and the runtime’s working headroom. Context memory is computed from each model’s declared attention pattern layer by layer, not from the generic formula — which is why several models fit here that other calculators say do not.
| Model | Compression shown | Model size | 8k | 32k | 128k | Largest context that fits |
|---|---|---|---|---|---|---|
| qwen3-1.7b | Q4_K_M |
1.03 GiB | yes | yes | — | 32k |
| qwen3-4b-2507 | Q4_K_M |
2.33 GiB | yes | yes | yes | 128k |
| gemma4-e2b | Q4_K_M |
2.89 GiB | yes | yes | yes | 128k |
| gemma4-e4b | Q4_K_M |
4.63 GiB | yes | yes | yes | 128k |
| qwen3-8b | Q4_K_M |
4.68 GiB | yes | yes | — | 32k |
| gemma4-12b | Q4_K_M |
6.63 GiB | yes | yes | yes | 256k |
| gpt-oss-20b | Q4_K_M |
10.83 GiB | yes | yes | yes | 128k |
| mistral-small-24b | Q4_K_M |
13.35 GiB | yes | yes | no | 64k |
| qwen3.8-27b | Q4_K_M |
15.33 GiB | yes | yes | yes | 128k |
| qwen3.6-27b | Q4_K_M |
15.66 GiB | yes | yes | yes | 128k |
| gemma4-26b-a4b | Q4_K_M |
15.78 GiB | yes | yes | yes | 128k |
| gemma4-31b | Q4_K_M |
17.07 GiB | yes | yes | no | 32k |
| qwen3-coder-30b | Q4_K_M |
17.28 GiB | yes | yes | no | 64k |
| qwen3.6-35b-a3b | Q4_K_M |
20.61 GiB | yes | yes | yes | 128k |
| gpt-oss-120b | Q4_K_M |
58.46 GiB | no | no | no | — |
| llama4-scout-17b | Q4_K_M |
60.87 GiB | no | no | no | — |
| deepseek-v4-flash | IQ4_XS |
127.28 GiB | no | no | no | — |
| glm-5.2 | Q4_K_M |
433.83 GiB | no | no | no | — |
Sizes are the real published files at the reference compression, one step per row: Q4_K_M where the author publishes it, the nearest neighbour where they do not. Every row states which one it is showing. Q4_K_M is our quality floor: below it the loss is audible in the answers. A model that would only fit here at a harsher compression is listed as not fitting, on purpose. A dash in a context column means the model itself does not offer that context, so there is nothing to fit.
Models that fit — 14 of the 18 models measured
- qwen3-1.7b
Q4_K_M— context 32k tokens - qwen3-4b-2507
Q4_K_M— context 128k tokens - gemma4-e2b
Q4_K_M— context 128k tokens - gemma4-e4b
Q4_K_M— context 128k tokens - qwen3-8b
Q4_K_M— context 32k tokens - gemma4-12b
Q4_K_M— context 256k tokens - gpt-oss-20b
Q4_K_M— context 128k tokens - mistral-small-24b
Q4_K_M— context 64k tokens - qwen3.8-27b
Q4_K_M— context 128k tokens - qwen3.6-27b
Q4_K_M— context 128k tokens - gemma4-26b-a4b
Q4_K_M— context 128k tokens - gemma4-31b
Q4_K_M— context 32k tokens - qwen3-coder-30b
Q4_K_M— context 64k tokens - qwen3.6-35b-a3b
Q4_K_M— context 128k tokens
Models that do not fit — 4 of the 18 models measured
- gpt-oss-120b
Q4_K_M— short by 30.28 GiB - llama4-scout-17b
Q4_K_M— short by 33.30 GiB - deepseek-v4-flash
IQ4_XS— short by 99.29 GiB - glm-5.2
Q4_K_M— short by 420.13 GiB