DGX Spark (GH200 480GB) vs Tesla P4 8GB for Local LLMs

The DGX Spark (GH200 480GB) has about 160% more memory bandwidth (500 vs 192 GB/s). Batch-1 decode tracks bandwidth closely, so a model that fits both cards should decode roughly 160% faster on the DGX Spark (GH200 480GB). The DGX Spark (GH200 480GB) fits more of the catalog at Q4/8K — 66 vs 16 of 69 models (480 vs 8 GB VRAM). Largest model that fits the DGX Spark (GH200 480GB) but not the Tesla P4 8GB: DeepSeek-V4-Pro (861.6B). MSRP is $3,000 for the DGX Spark (GH200 480GB) vs $80 for the Tesla P4 8GB. MSRP is a launch list price, not live retail — Amazon prices move; click through for the current price. In the Q4/8K head-to-head below (39 shared models), the DGX Spark (GH200 480GB) posts the higher tok/s in 39 rows vs 0. Methodology.

Spec comparison

	DGX Spark (GH200 480GB)	Tesla P4 8GB
Vendor	NVIDIA	NVIDIA
VRAM	480 GB	8 GB
Memory type	LPDDR5X	GDDR5
Memory bandwidth	500 GB/s	192 GB/s
TDP	170 W	75 W
MSRP (list)	$3,000	$80
Buy	Search Amazon (affiliate)	Search Amazon (affiliate)

MSRP is the launch list price, not live retail. Buy links are Amazon affiliate links — the price after you click is the live price.

Q4 / 8K head-to-head — top 15 of 39 shared models

Model	Params (B)	DGX Spark (GH200 480GB) tok/s	DGX Spark (GH200 480GB) fits	Tesla P4 8GB tok/s	Tesla P4 8GB fits	Winner
MiniCPM5-1B	1.08	294.6 est.	Yes	113.1 est.	Yes	DGX Spark (GH200 480GB)
Qwen3.5-2B	2.0	159.1 est.	Yes	61.1 est.	Yes	DGX Spark (GH200 480GB)
DeepSeek-R1-0528-Qwen3-8B-layer-mix-bpw-3.8-mlx	2.3	138.3 est.	Yes	53.1 est.	Yes	DGX Spark (GH200 480GB)
Qwen3.6-35B-A3B	36.0	106.1 est.	Yes	17.0 est.	No	DGX Spark (GH200 480GB)
Qwen3.5-35B-A3B	34.7	106.1 est.	Yes	18.2 est.	No	DGX Spark (GH200 480GB)
GLM-4.7-Flash	31.2	106.1 est.	Yes	21.0 est.	No	DGX Spark (GH200 480GB)
Qwen3-30B-A3B-Instruct-2507	30.5	106.1 est.	Yes	21.9 est.	No	DGX Spark (GH200 480GB)
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16	30.0	106.1 est.	Yes	24.1 est.	No	DGX Spark (GH200 480GB)
Ornith-1.0-35B	35.0	106.1 est.	Yes	17.0 est.	No	DGX Spark (GH200 480GB)
qwen35b-a3b-fable-sft-abliterated	34.7	106.1 est.	Yes	17.3 est.	No	DGX Spark (GH200 480GB)
Llama-3.2-3B-Instruct	3.2	99.4 est.	Yes	38.2 est.	Yes	DGX Spark (GH200 480GB)
gpt-oss-20b	21.5	88.4 est.	Yes	33.6 est.	No	DGX Spark (GH200 480GB)
gemma-4-26B-A4B-it	25.2	83.7 est.	Yes	23.9 est.	No	DGX Spark (GH200 480GB)
Qwen3-4B-Instruct-2507	4.0	79.5 est.	Yes	30.5 est.	Yes	DGX Spark (GH200 480GB)
NVIDIA-Nemotron-3-Nano-4B-BF16	4.0	79.5 est.	Yes	30.5 est.	Yes	DGX Spark (GH200 480GB)

est. = bandwidth-based estimate; measured = curated real benchmark that overrides the estimate. Fits = weights + KV(8K) + 1 GB overhead ≤ 95% of VRAM.

Fit coverage at Q4 / 8K

	DGX Spark (GH200 480GB)	Tesla P4 8GB
Catalog models that fit	66 of 69	16 of 69
Largest exclusive fit	DeepSeek-V4-Pro (861.6B)	—

DGX Spark (GH200 480GB) details → · Tesla P4 8GB details →