More
Best rig by budgetRent before you buySubmit a buildNewsletterFree datasetAdvertise
Submit build

Builds / 4x Nvidia GB10 (unified, DGX Spark)

4x DGX Spark: DeepSeek V4.1 Flash EXL3 3.5 bpw on TensorFold TP4, 78 tok/s prose (Tech2wild)

Last updated:

Source: github.com ↗More source: github.com ↗📏 How measured: reported by the original source; RigBase re-checked the numbers at that source
GPU(s)
4x Nvidia GB10 (unified, DGX Spark)
VRAM / unified memory (GB)
512
CPU
Nvidia GB10 Grace (Arm) per node
Amazon · eBay (used) · Newegg
Motherboard
not stated
RAM
128 GB LPDDR5X unified per Spark (512 GB total)
PSU
not stated
Other
4 NVIDIA DGX Sparks, multi-node
Reported total cost
not stated
Power
not stated

Measured speed (generation tokens/sec)

ModelQuantEnginetok/sPrompt processing (tok/s)Details
DeepSeek V4.1 FlashEXL3 3.5 bpw + DSparkTensorFold TP4, 4 nodes, 1 stream78not stated⚡ spec-decode (DSpark)context: 1M📏 Single-stream prose decode as published in Tech2wild's recipe README (bertholomus/deepseek-v4.1-tensorfold-tp4-4xgb10 (3.5 bpw lane; counting excluded)).
DeepSeek V4.1 FlashEXL3 3.5 bpw + DSparkTensorFold TP4, 4 nodes, 1 stream134.8not stated⚡ spec-decode (DSpark)context: 1M📏 Single-stream code decode as published in the same Tech2wild README.

Numbers by Tech2wild (tonyd2wild); bertholomus/deepseek-v4.1-tensorfold-tp4-4xgb10 (3.5 bpw lane; counting excluded). RigBase read the GitHub README on 2026-10-28 and checked each figure against the text. RigBase did not re-run it. Speculative decoding (MTP/DFlash/DSpark) inflates predictable text; these rows prefer prose/code over counting peaks. Recipe license: MIT.

Source: https://github.com/tonyd2wild/DeepSeek-V4.1-Flash-TP2-and-TP4-DGX-Sparks
https://github.com/tonyd2wild/DeepSeek-V4.1-Flash-TP2-and-TP4-DGX-Sparks/blob/main/README.md

Blank fields = not stated in the source. We never estimate build data; GPU street prices above are separately sourced and dated.

Embed this result

Copy into a README or blog. Links back to this build. See badge docs (CC BY 4.0 attribution). The SVG does not invent speeds — numbers stay on the build page.

Markdown

[![Measured on RigBase](https://getrigbase.com/badge/tony-4x-dgx-spark-deepseek-v4-1-flash-tp2-and-tp4-dgx-sparks.svg)](https://getrigbase.com/builds/tony-4x-dgx-spark-deepseek-v4-1-flash-tp2-and-tp4-dgx-sparks/)

HTML

<a href="https://getrigbase.com/builds/tony-4x-dgx-spark-deepseek-v4-1-flash-tp2-and-tp4-dgx-sparks/" rel="noopener"><img src="https://getrigbase.com/badge/tony-4x-dgx-spark-deepseek-v4-1-flash-tp2-and-tp4-dgx-sparks.svg" alt="Measured on RigBase" height="20"/></a>

Preview: Measured on RigBase · How to use

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 4x DGX Spark (GB10) run DeepSeek V4.1 Flash?