Rig score · nvidia
Nvidia DGX Spark (128GB)
Runs 143 of 161 models (89%)
Elite, runs the biggest open models, including 70B.
Biggest it can run: dots.llm1
- Grade
- S
- Runnable
- 143
- Total
- 161
- Coverage
- 89%
What Nvidia DGX Spark (128GB) can run
At 128 GB of unified memory, Nvidia DGX Spark (128GB) runs 143 of 161 tracked local models, 89% coverage that earns a S. The largest it handles is dots.llm1 at 142B parameters; anything bigger needs more memory or a lower quantization.
CPU and GPU share one 128 GB pool, so there is no separate VRAM to spill out of. Memory bandwidth sets the speed: mixture-of-experts models run well here, while large dense models load but generate slowly.
Top models on Nvidia DGX Spark (128GB)
FAQ
What local AI rig score does Nvidia DGX Spark (128GB) get?
Nvidia DGX Spark (128GB) scores S: it runs 143 of 161 tracked local LLMs (89%). The biggest it can run is dots.llm1.
How big a model can Nvidia DGX Spark (128GB) run?
Nvidia DGX Spark (128GB) runs models up to dots.llm1 (142B parameters) at Q4_K_M. Its memory is shared with the system, so leave a few GB for the OS.
Is Nvidia DGX Spark (128GB) good for running local LLMs?
CPU and GPU share one 128 GB pool, so there is no separate VRAM to spill out of. Memory bandwidth sets the speed: mixture-of-experts models run well here, while large dense models load but generate slowly.
Want your own? Check your rig on the homepage →
Score from coverage of 161 tracked models, estimated with our memory methodology.