Llama on NVIDIA Blackwell

Can I run Llama 3.2 1B on an NVIDIA DGX Spark (GB10)?

Yes. Llama 3.2 1B at Q4_K_M uses 1.5 GB of the 96.0 GB available on a NVIDIA DGX Spark (GB10), and runs at about 209 tokens per second. You can push the context to 128K.

Runs comfortably

1.5 GB of 96.0 GB · 2%
096 GB
Weights 0.7 GB
KV cache 0.3 GB
Runtime overhead 0.5 GB

Llama 3.2 1B at Q4_K_M leaves 94.5 GB spare on a NVIDIA DGX Spark (GB10). There is room to raise the context length or move up a quantisation level.

Generation209tok/s
Prompt processing21169tok/s
Max context128Ktokens
KV per 1K tokens0GB

Every quantisation of Llama 3.2 1B on a NVIDIA DGX Spark (GB10)

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 2.3 GB 3.0 GB Runs comfortably 128K 73.6
INT8 / W8A8 8.50 1.2 GB 1.9 GB Runs comfortably 128K 138
Q8_0 (GGUF) 8.50 1.2 GB 1.9 GB Runs comfortably 128K 138
FP8 (E4M3) 8.00 1.2 GB 1.9 GB Runs comfortably 128K 139
AWQ 4-bit 4.25 1.0 GB 1.7 GB Runs comfortably 128K 162
GPTQ 4-bit 4.25 1.0 GB 1.7 GB Runs comfortably 128K 162
MXFP4 4.25 1.0 GB 1.7 GB Runs comfortably 128K 162
Q6_K 6.56 0.9 GB 1.7 GB Runs comfortably 128K 166
Q5_K_M 5.67 0.8 GB 1.6 GB Runs comfortably 128K 188
Q5_K_S 5.52 0.8 GB 1.5 GB Runs comfortably 128K 192
Q4_K_M 4.85 0.7 GB 1.5 GB Runs comfortably 128K 209
Q4_K_S 4.58 0.7 GB 1.4 GB Runs comfortably 128K 217
Q4_0 4.55 0.7 GB 1.4 GB Runs comfortably 128K 218
IQ4_XS 4.25 0.7 GB 1.4 GB Runs comfortably 128K 227
Q3_K_M 3.91 0.6 GB 1.3 GB Runs comfortably 128K 239
IQ3_M 3.70 0.6 GB 1.3 GB Runs comfortably 128K 247
IQ3_XXS 3.06 0.5 GB 1.3 GB Runs comfortably 128K 274
Q2_K 2.63 0.5 GB 1.2 GB Runs comfortably 128K 296
IQ2_XXS 2.06 0.4 GB 1.1 GB Runs comfortably 128K 331
IQ1_M 1.75 0.4 GB 1.1 GB Runs comfortably 128K 354

Also worth checking

Llama 3.2 1B on every GPU
Full hardware tableEverything that runs on a NVIDIA DGX Spark (GB10)
Full model tableWhy the KV cache is the part that bites
Guide