DGX Spark with Qwen3.6 27B NVFP4 + DFlash speculative decoding
| GPU(s) | Nvidia GB10 (unified) |
|---|---|
| VRAM / unified memory (GB) | 128 |
| CPU | Nvidia Grace (GB10) Amazon · eBay (used) · Newegg |
| Motherboard | |
| RAM | 128 GB unified |
| PSU | |
| Other | vLLM + z-lab DFlash drafter, thinking on |
| Reported total cost | |
| Power |
Measured speed (generation tokens/sec)
📏 How measured: reported by the original source; RigBase re-checked the numbers at that source
| Model | Quant | Engine | tok/s | Details |
|---|---|---|---|---|
| Qwen3.6 27B | NVFP4 | vLLM + DFlash spec decoding | 32.83 | ⚡ spec-decode (DFlash)📏 Median decode, Docker reproduction |
| Qwen3.6 27B | FP8 | vLLM autoregressive baseline | 7.85 | 📏 Autoregressive baseline, same reproduction |
Docker reproduction; median decode 32.83 tok/s (mean 40.10). Thinking off drops speed sharply. Found via localmaxxing (discovery only); numbers re-checked at the original source on 2026-10-24.
Source: https://github.com/my-other-github-account/qwen36-dflash-spark-repro
Blank fields = not stated in the source. We never estimate build data; GPU street prices above are separately sourced and dated.
Rent before you buy: try this model on a cloud GPU for a few dollars first: Runpod (referral link; sign up with Google to get the new-user bonus)
Quick answers: Can DGX Spark (GB10) run Qwen3.6 27B?