Large model prototyping (up to 200B)
NVIDIA DGX Spark
NVIDIA · Verified 2026-07-23
From $3,999
Best for developers who need 128GB unified memory for large model prototyping without multi-GPU complexity.
Main limitation: ARM-based, no Windows support, not user-upgradeable
Some links are affiliate links. We may earn a commission at no extra cost to you. Full disclosure.
Key Specifications
| Specification | Value |
|---|---|
| Form Factor | compact desktop |
| GPU | NVIDIA Blackwell (integrated in GB10) |
| GPU Count | 1 |
| VRAM per GPU | 128 |
| Total Accelerator Memory | 128 |
| Memory Architecture | unified (LPDDR5x) |
| Memory Bandwidth (per GPU) | 273 |
| Unified Memory | 128 |
| System RAM | 128 |
| RAM Type | LPDDR5x unified |
| CPU | NVIDIA Grace Blackwell (GB10) |
| CPU Cores | 20 (10 Cortex-X925 + 10 Cortex-A725) |
| Primary Storage | 4096 |
| Storage Type | NVMe SSD |
| Power Supply | 150 |
| Est. Peak Power | 150 |
| Cooling | active air |
| Dimensions | 202 × 202 × 63 mm |
| Weight | 2.3 |
| OS | NVIDIA DGX OS (Ubuntu-based) |
| Linux Support | Yes (DGX OS) |
| Windows Support | No |
| WSL Support | No |
| CUDA Support | Yes |
| Metal Support | No |
| Warranty | 3 |
| Onsite Support | No |
| Release Year | 2025 |
What Models Can It Run?
Calculated estimates using weight-only memory analysis at Q4 (4-bit) quantization. Actual requirements vary by framework, context length, and runtime overhead. See methodology.
| Model | Params | Required VRAM (Q4) | Fits in 94.5GB? |
|---|---|---|---|
| Llama 3.1 8B | 8B | 5.52 GB | ✓ Yes |
| Llama 3.1 70B | 70B | 37.14 GB | ✓ Yes |
| Llama 3.2 1B | 1B | 2 GB | ✓ Yes |
| Llama 3.2 3B | 3B | 3.01 GB | ✓ Yes |
| Qwen 2.5 7B | 7B | 5.01 GB | ✓ Yes |
| Qwen 2.5 14B | 14B | 8.53 GB | ✓ Yes |
| Qwen 2.5 32B | 32B | 18.06 GB | ✓ Yes |
| Qwen 2.5 72B | 72B | 38.14 GB | ✓ Yes |
| Mistral 7B | 7B | 5.01 GB | ✓ Yes |
| Mixtral 8x7B | 46.7B | 25.44 GB | ✓ Yes |
| DeepSeek R1 7B | 7B | 5.01 GB | ✓ Yes |
| DeepSeek R1 32B | 32B | 18.06 GB | ✓ Yes |
| DeepSeek R1 70B | 70B | 37.14 GB | ✓ Yes |
| Phi-3 Medium 14B | 14B | 8.53 GB | ✓ Yes |
| Gemma 2 9B | 9B | 6.02 GB | ✓ Yes |
| Gemma 2 27B | 27B | 15.05 GB | ✓ Yes |
| Stable Diffusion XL | 6.6B | 4.81 GB | ✓ Yes |
| Flux.1 Dev | 12B | 7.52 GB | ✓ Yes |
| Flux.1 Schnell | 12B | 7.52 GB | ✓ Yes |
Max model size: 70B class models (~182B params at Q4). These are calculated estimates, not measured results.
Pros & Cons
Pros
- ✓ Very large accelerator memory (128GB)
- ✓ Good memory per dollar ($31/GB)
- ✓ Linux support
- ✓ 3-year warranty
Cons
- ✗ ARM-based, no Windows support, not user-upgradeable
Compare with Other Workstations
Apple Mac Studio M2 Ultra (192GB)
Apple Mac Studio M2 Ultra (64GB)
ArsenalPC MES2X Dual RTX 5090 AI Workstation
BIZON G3000 G2 (1x RTX 5090)
BIZON G3000 G2 (4x RTX 5090)
BOXX APEXX 8R (1x RTX PRO 6000 Blackwell)
Custom Dual RTX 4090 Training Workstation
Custom RTX 5090 AI Workstation (Value Build)
Sources & Verification
Primary source: NVIDIA DGX Spark Official Page
https://www.nvidia.com/en-us/products/workstations/dgx-spark/
Verification date: 2026-07-23
Data quality: Vendor specification — from manufacturer datasheet. No hands-on testing was performed.