NVIDIA RTX Spark N1X (128 GB)
Popular models on the NVIDIA RTX Spark N1X (128 GB)
Fit at Q4_K_M and estimated speed on 128 GB
| Model | Q4 size | Fit | Est. speed |
|---|---|---|---|
| gpt-oss-120b | 70 GB | Fits | ~35 tok/s |
| Qwen2.5 72B | 40 GB | Fits | ~3.9 tok/s |
| Llama 3.3 70B | 40 GB | Fits | ~3.9 tok/s |
| Qwen3.6 35B A3B | 21 GB | Fits | ~60 tok/s |
| Qwen3.5 35B A3B | 21 GB | Fits | ~60 tok/s |
| dolphin-2.9.1-yi-1.5-34b | 20 GB | Fits | ~7.8 tok/s |
| Gemma 4 31B | 19 GB | Fits | ~8.3 tok/s |
| Qwen3 32B | 19 GB | Fits | ~8.3 tok/s |
| GLM-4.7-Flash | 18 GB | Fits | ~60 tok/s |
| Qwen3-30B-A3B-Instruct-2507 | 18 GB | Fits | ~59 tok/s |
| Qwen3.6 27B | 16 GB | Fits | ~9.8 tok/s |
| Qwen3.8 27B | 16 GB | Fits | ~9.8 tok/s |
Estimates: Q4_K_M weights + 1.8 GB overhead; speed at Q4_K_M, scaled from memory bandwidth.
Local inference speed
llama.cpp decode speed (tok/s) - estimated
| Workload | tok/s |
|---|---|
| LLaMA 3 8B Q4 | 32 |
| LLaMA 3 8B F16 | 14.5 |
| LLaMA 3 70B Q4 | 3.8 |
Estimated - Estimated from memory bandwidth (llama.cpp decode); bandwidth assumed equal to DGX Spark (GB10), which shares the 6,144-core GPU - NVIDIA has not published an N1X figure · updated 2026-11-01
Stacks for the NVIDIA RTX Spark N1X (128 GB)
Complete local setups that run on 128 GB
Run Qwen3.6-27B with vision, 256K context, and hybrid thinking locally
AI pair programming in your terminal with local models, fully private
A private AI assistant on Windows with document RAG - no CLI, no Docker, all desktop GUI
A private AI coding assistant in your editor
Prototype and deploy multi-agent RAG applications visually with local models
NVIDIA RTX Spark N1X (128 GB)Link to this section
The RTX Spark N1X is NVIDIA's first Windows PC chip: a 20-core Grace (Arm) CPU and a 6,144-core Blackwell RTX GPU sharing up to 128 GB of LPDDR5X unified memory, rated up to 1 petaflop of FP4 AI compute. Laptops started shipping in mid-October 2026 (ASUS ProArt P16, Dell XPS 16, HP OmniBook Ultra 16, Lenovo Pro 9n, Microsoft Surface Laptop Ultra, MSI Prestige N16 Flip AI+), at 45-80 W.
For local AI it is the same recipe as the DGX Spark in a laptop: lots of memory, modest bandwidth. Memory capacity decides which models fit; bandwidth sets how fast they generate.
What 128 GB means hereLink to this section
- It's unified memory, not 128 GB of VRAM. CPU and GPU share it, and NVIDIA says the GPU can address less than the total installed, depending on configuration. Budget roughly 100-110 GB for model weights plus context.
- Only the top configuration has 128 GB. Retail models start at 24-32 GB; check the exact SKU before buying for local LLMs.
- Big mixture-of-experts models are the sweet spot. They need lots of memory but read only their active parameters per token, so they run far faster than dense models of the same size on ~273 GB/s.
SpeedLink to this section
NVIDIA hasn't published the N1X memory bandwidth. The estimates on this page assume the 256-bit LPDDR5X bus of the DGX Spark's GB10 (273 GB/s), which uses the same GPU configuration; third-party reports range from 273 to 300 GB/s. Treat speeds as estimates until measured results are out.
Check any model with the VRAM calculator, or see everything that fits on what can I run.
Frequently asked
Quick answers to common questions
How much VRAM does the NVIDIA RTX Spark N1X (128 GB) have?
The NVIDIA RTX Spark N1X (128 GB) has 128 GB of VRAM with 273 GB/s memory bandwidth.
What local AI models can run on the NVIDIA RTX Spark N1X (128 GB)?
The NVIDIA RTX Spark N1X (128 GB) with 128 GB VRAM can run many models depending on quantization. Models up to ~196B params may fit at Q4_K_M. Use our VRAM calculator to check specific models.
Is the NVIDIA RTX Spark N1X (128 GB) good for local AI inference?
NVIDIA RTX Spark N1X (128 GB) is best for llm-inference, image-gen, comfyui. With ample VRAM it handles most open models well.
Where can I buy the NVIDIA RTX Spark N1X (128 GB)?
Check our buy links above for the best current prices on Amazon, Newegg, and B&H. Prices vary by retailer and availability.
How does the NVIDIA RTX Spark N1X (128 GB) compare to other GPUs?
NVIDIA RTX Spark N1X (128 GB) has 128 GB VRAM and 273 GB/s bandwidth. This puts it in the high-end category, suitable for most open models. Browse our hardware directory for side-by-side comparisons.
What power supply do I need for the NVIDIA RTX Spark N1X (128 GB)?
The NVIDIA RTX Spark N1X (128 GB) has a TDP of 80W. A standard quality PSU of 650W+ should suffice. Always check the manufacturer's recommendations for your specific build.
Nearby options
Similar hardware and models that fit