Can Qwen3.5 27B run on the DGX Spark?
Yes. At 8K context the best fit is F16, using about 52.3 GB of the 109 GB usable on the DGX Spark. Estimated speed ceiling: ~5 tokens/sec.
+ 0.9 GB with image input. The figures on this page are for text only. With image input enabled, the vision projector (mmproj) file loads alongside the model: F16 at 8K context then needs about 53.1 GB (still fits). llama.cpp's -hf downloads and loads it automatically unless you add --no-mmproj. Calculator with image input on.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | 12.3 GB Fits | 12.6 GB Fits | 14.1 GB Fits | 20.1 GB Fits | 23 |
| 3-bit (Q3_K_M) • | 14.9 GB Fits | 15.2 GB Fits | 16.7 GB Fits | 22.7 GB Fits | 18 |
| 4-bit (Q4_K_M) • | 17.9 GB Fits | 18.1 GB Fits | 19.6 GB Fits | 25.6 GB Fits | 15 |
| 5-bit (Q5_K_M) • | 20.6 GB Fits | 20.9 GB Fits | 22.4 GB Fits | 28.4 GB Fits | 13 |
| 6-bit (Q6_K) • | 23.2 GB Fits | 23.4 GB Fits | 24.9 GB Fits | 30.9 GB Fits | 12 |
| 8-bit (Q8_0) • | 28.2 GB Fits | 28.5 GB Fits | 30.0 GB Fits | 36.0 GB Fits | 9 |
| 16-bit (F16/BF16) • | 52.0 GB Fits | 52.3 GB Fits | 53.8 GB Fits | 59.8 GB Fits | 5 |
• real GGUF file size from bartowski/Qwen_Qwen3.5-27B-GGUF. Speed is a bandwidth ceiling (273 GB/s), not a benchmark. Methodology.
What fits instead
- Cheapest listed GPU that runs Qwen3.5 27B: RX 7900 XTX (24 GB, launch MSRP $999) at Q4_K_M, fully in VRAM at 8K context.
- Largest popular model that fits the DGX Spark: gpt-oss 120B (117B, MoE) at F16, about 61.9 GB at 8K context.
"Runs" and "fits" mean the whole model in GPU memory at 8K context (4-bit or better for the model pick). Prices are US launch MSRPs, not today's street prices. "Popular" means the 60 most-downloaded models we track.
More
Qwen3.5 27B on every GPU and Mac · Every model that runs on the DGX Spark · Other models that fit 64 GB VRAM · Try other settings in the calculator
Guides: Best GPU for local LLMs by budget · Can I run gpt-oss, Llama 70B or Qwen3? · Mac vs PC for local LLMs
Data: Hugging Face Hub (Qwen/Qwen3.5-27B), NVIDIA specs. Model data last checked 2026-11-18; refreshed hourly.