readygpu

NVIDIA DGX Spark 128GB vs NVIDIA GeForce RTX 5090

For running models locally, priced in the UK. At 64k context the first runs 75 models fully in memory, the second 54.

In short

The NVIDIA DGX Spark 128GB has more memory for models: 128 GB against 32 GB. At 64k context the NVIDIA DGX Spark 128GB runs 75 of the models we track fully in memory and the NVIDIA GeForce RTX 5090 54; only the NVIDIA DGX Spark 128GB holds DeepSeek V4 Flash. On the models both run fully in memory, the NVIDIA GeForce RTX 5090 is about 6.6× as fast as the NVIDIA DGX Spark 128GB (estimated, median across models). In the UK the NVIDIA DGX Spark 128GB costs £5,499 at Amazon and the NVIDIA GeForce RTX 5090 £5,400 at eBay: the NVIDIA GeForce RTX 5090 is £99 cheaper.

Updated . Speeds are estimates; how they are worked out.

NVIDIA DGX Spark 128GB

£5,499

Amazon

Add to build
NVIDIA GeForce RTX 5090

£5,400

eBay

Add to build

Specifications

SpecNVIDIA DGX Spark 128GBNVIDIA GeForce RTX 5090
ChipGB10 Grace Blackwell–
Memory bus256-bit512-bit
Memory bandwidth273 GB/s1,792 GB/s
FP16 compute62 TFLOPS209 TFLOPS
SoftwareCUDACUDA
Unified memory128 GB–
GPU can use by default119 GB–
GPU can use, tuned122 GB–
Operating systemdgx-os–
Processor20-core Arm (GB10)–
Storage4000 GB–
GPU–GeForce RTX 5090
Memory–32 GB GDDR7
Board power–575 W
PCIe generation–Gen 5
PCIe lanes–x16

What each runs

Estimated decode speed with the most precise quant that fits each one est.

ModelNVIDIA DGX Spark 128GBNVIDIA GeForce RTX 5090
DeepSeek V4 Flash8tok/sUD-Q2_K_XLwon't run
GLM 5.3 Flash11tok/sUD-Q2_K_XLwon't run
Qwen 3.8 Flash Next2tok/sQ4_K_Mwon't run
Qwen 3.8 27B5tok/sQ8_039tok/sQ6_K
MiniMax M2.75tok/sUD-Q3_K_Mwon't run
Qwen 3.5 122B-A10B10tok/sQ6_Koffload
Qwen 3.6 27B5tok/sQ8_041tok/sQ6_K
gpt-oss 120B18tok/sMXFP4won't run
Qwen3 Coder Next18tok/sQ8_0offload
Nemotron 3 Super 120B-A12B9tok/sUD-Q6_Koffload
Gemma 4 31B4tok/sQ8_035tok/sQ6_K
Mistral Small 4 119B14tok/sUD-Q6_Koffload
Qwen 3.6 35B-A3B23tok/sQ8_0167tok/sQ6_K
Qwen 3.5 27B5tok/sQ8_041tok/sQ6_K
Gemma 4 26B-A4B18tok/sQ8_0117tok/sQ8_0
GLM 4.5 Air4tok/sQ6_Koffload
Qwen 3.5 35B-A3B23tok/sQ8_0175tok/sQ6_K
Qwen3 Next 80B-A3B21tok/sQ8_0offload
GLM 4.7 Flash12tok/sQ8_087tok/sQ6_K
Devstral Small 2 24B5tok/sQ8_037tok/sQ6_K
Nemotron 3.5 Lightning 30B-A3B27tok/sUD-Q6_K_XL201tok/sUD-Q5_K_M
Kimi Linear 48B-A3B27tok/sQ8_0277tok/sQ4_K_M
gpt-oss 20B25tok/sMXFP4165tok/sMXFP4
Qwen3 Coder 30B-A3B11tok/sQ8_075tok/sQ6_K
Qwen3 32B3tok/sQ8_0offload
Gemma 4 12B12tok/sQ8_079tok/sQ8_0
Qwen 3.5 9B14tok/sQ8_095tok/sQ8_0
Gemma 3 27B5tok/sQ8_040tok/sQ6_K
Mistral Small 3.2 24B5tok/sQ8_037tok/sQ6_K
DeepSeek R1 Distill Qwen 32B3tok/sQ8_0offload
Llama 3.3 70B2tok/sQ8_0won't run
Ministral 3 14B7tok/sQ8_044tok/sQ8_0
Qwen3 14B6tok/sQ8_042tok/sQ8_0
DeepSeek R1 0528 Qwen3 8B9tok/sQ8_061tok/sQ8_0
Qwen3 8B9tok/sQ8_061tok/sQ8_0
Phi-4 14B6tok/sQ8_038tok/sQ8_0
Qwen 3.5 4B25tok/sQ8_0166tok/sQ8_0
Ministral 3 8B9tok/sQ8_061tok/sQ8_0
Granite 4.2 8B8tok/sQ8_055tok/sQ8_0
Gemma 4 E4B17tok/sQ8_0110tok/sQ8_0
LFM2.5 8B-A1B24tok/sQ8_0156tok/sQ8_0
Llama 3.1 8B11tok/sQ6_K73tok/sQ6_K
Qwen2.5 0.5B127tok/sQ8_0831tok/sQ8_0
Qwen3.5 0.8B104tok/sQ8_0679tok/sQ8_0
Qwen3.5 2B75tok/sQ5_K_M492tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B45tok/sQ8_0294tok/sQ8_0
MiniCPM5 2B31tok/sQ8_0202tok/sQ8_0
Spark X2.5 4B31tok/smlx-4bit202tok/smlx-4bit
Mellum2.1 12B A2.5B Thinking28tok/sQ8_0183tok/sQ8_0
LFM2.5 2.6B25tok/sQ8_0161tok/sQ8_0
LFM2.5 VL 3B25tok/sQ8_0161tok/sQ8_0
Kimi VL A3B Thinking 250620tok/sQ8_0130tok/sQ8_0
Kimi VL A3B19tok/smlx-8bit126tok/smlx-8bit
granite vision 4.1 4b19tok/sQ8_0124tok/sQ8_0
granite 4.2 3b18tok/sQ8_0120tok/sQ8_0
GLM 4 9B 041416tok/smlx-5bit102tok/smlx-5bit
DeepSeek R1 Distill Qwen 7B14tok/sQ8_094tok/sQ8_0
LFM2.5 24B A2B14tok/sQ8_089tok/sQ8_0
GLM 4.6V Flash13tok/sQ8_088tok/sQ8_0
GLM 4.1V 9B Thinking12tok/smlx-8bit77tok/smlx-8bit
DeepSeek R1 Distill Llama 8B10tok/sQ8_065tok/sQ8_0
Ministral 3 8B Reasoning 25129tok/sQ8_061tok/sQ8_0
Qwen3.8 27B OBLITERATED7tok/smlx-4bit44tok/smlx-4bit
DeepSeek R1 Distill Qwen 14B6tok/sQ8_039tok/sQ8_0
Devstral Small 25075tok/sQ8_040tok/sQ5_K_M
Magistral Small 25065tok/sQ8_037tok/sQ6_K
Magistral Small 25075tok/sQ8_040tok/sQ5_K_M
Magistral Small 25095tok/sQ8_040tok/sQ5_K_M
GLM 4.5V4tok/sQ8_0won't run
GLM 4.6V4tok/sQ8_0won't run
granite 4.2 30b4tok/sQ8_0offload
DeepSeek R1 Distill Llama 70B2tok/sQ8_0won't run
Kimi Dev 72B2tok/sQ8_0won't run
Devstral 2 123B Instruct 25121tok/sQ6_Kwon't run
Mistral Medium 3.5 128B1tok/sQ6_Kwon't run
DeepSeek R1won't runwon't run
DeepSeek R1 0528won't runwon't run
DeepSeek R1 Zerowon't runwon't run
DeepSeek V3 0324won't runwon't run
DeepSeek V3.1won't runwon't run
DeepSeek V3.1 Terminuswon't runwon't run
DeepSeek V3.2won't runwon't run
DeepSeek V4 Flash 0731won't runwon't run
DeepSeek V4 Flash Vision Expwon't runwon't run
GLM 4.6won't runwon't run
GLM 4.7won't runwon't run
GLM 5.1won't runwon't run
GLM 5.2won't runwon't run
GLM 5.3won't runwon't run
Kimi K2won't runwon't run
Kimi K2 Instruct 0905won't runwon't run
Kimi K2 Thinkingwon't runwon't run
Kimi K2.5won't runwon't run
MiniMax M2won't runwon't run
MiniMax M2.1won't runwon't run
MiniMax M2.5won't runwon't run
MiniMax M3won't runwon't run

Questions

Which is better for local AI, the NVIDIA DGX Spark 128GB or the NVIDIA GeForce RTX 5090?
Choose the NVIDIA DGX Spark 128GB to run bigger models. Choose the NVIDIA GeForce RTX 5090 for speed. Choose the NVIDIA GeForce RTX 5090 to spend less.
Which can run bigger models, the NVIDIA DGX Spark 128GB or the NVIDIA GeForce RTX 5090?
At 64k context the NVIDIA DGX Spark 128GB runs 75 of the models we track fully in memory and the NVIDIA GeForce RTX 5090 54; only the NVIDIA DGX Spark 128GB holds DeepSeek V4 Flash.
Which is faster for local AI, the NVIDIA DGX Spark 128GB or the NVIDIA GeForce RTX 5090?
On the models both run fully in memory, the NVIDIA GeForce RTX 5090 is about 6.6× as fast as the NVIDIA DGX Spark 128GB (estimated, median across models).
What is the price difference between the NVIDIA DGX Spark 128GB and the NVIDIA GeForce RTX 5090?
In the UK the NVIDIA DGX Spark 128GB costs £5,499 at Amazon and the NVIDIA GeForce RTX 5090 £5,400 at eBay: the NVIDIA GeForce RTX 5090 is £99 cheaper.

More comparisons