Best Local LLMs for the NVIDIA DGX Spark

Written by Jakub Rusinowski · Last updated July 21, 2026

Best all-round pick: GPT-oss 120B

The NVIDIA DGX Spark has 128 GB of VRAM, of which about 128 GB is available to a model. The largest model it can hold is Devstral-2 123B (123B, 75.1 GB at Q4_K_M).

Best models overall on the NVIDIA DGX Spark

ModelScoreMemory at Q4_K_MContextLicence
GPT-oss 120B9673.3 GB125KApache-2.0
Nemotron 70B Instruct94.743.4 GB125KLlama Community
Qwen 3.5 122B-A10B (MoE)94.274.5 GB125KApache 2.0
Qwen 3.5 122B-A10B92.574.5 GB128KApache-2.0
GLM-5.1 72B91.244.3 GB125KMIT
Command R+ (104B)91.263.6 GB125KCC-BY-NC

Best model by what you are doing

WorkloadRecommended on this GPU
CodingDevstral-2 123B (98.5)
Qwen 3.5 122B-A10B (MoE) (96.5)
Qwen3-Coder 80B-A3B (MoE) (96.2)
General assistantGPT-oss 120B (96)
Nemotron 70B Instruct (94.7)
Qwen 3.5 122B-A10B (MoE) (94.2)
ReasoningGLM-5.1 72B (98.4)
Qwen 3.5 122B-A10B (MoE) (97)
Gemma 4 27B ⭐ (95.9)
RAGLlama 4 Scout 17B (93.6)
Gemma 4 31B (92.8)
Llama 4.5 Scout (92.3)
AgentsTernary Bonsai 27B (89)
Qwen 3.5 122B-A10B (MoE) (87.5)
Nemotron 3 Super 120B-A12B (86.6)
VisionQwen 3.5 72B (96.9)
Llama 3.2 90B Vision Instruct (95.2)
Mistral Small 4 119B-A6.5B (93.1)

How these numbers are calculated

FAQ

What is the best LLM for the NVIDIA DGX Spark?

GPT-oss 120B is the strongest all-round pick that fits its 128 GB.

What is the largest model the NVIDIA DGX Spark can run?

Devstral-2 123B — 123B parameters, needing 75.1 GB at Q4_K_M.

How much of the NVIDIA DGX Spark's memory can a model actually use?

About 128 GB of its 128 GB, before the desktop and runtime overhead are accounted for.

Compatibility Checks for the NVIDIA DGX Spark

By Workload

More

← All GPUs | NVIDIA DGX Spark specs