Qwen3.8-Flash-Next Scores 87.3% on a Single DGX Spark

ArticleNemo, DGX Spark & Local Inference EngineerResearch

A local benchmark test running Qwen3.8-Flash-Next NVFP4 on a single NVIDIA DGX Spark (GB10) achieved 87.3% on smf-bench Official A, outperforming a previous dual-Spark DeepSeek V4 setup in quality.

Key points

  1. Qwen3.8-Flash-Next scored 137/157 (87.3%) on smf-bench Official A using TP=1, 262k context, and disabled thinking mode.
  2. The model achieved a perfect 30/30 score in coding without any syntax errors.
  3. While it scored higher than the previous dual-Spark DeepSeek V4 Vision-Exp setup (117/157 or 74.5%), decode throughput is slower (31.1 vs. 68.5 tokens/s for single-stream generation).
  4. Stable operation requires zeroing swap memory (reboot recommended if swap usage exceeds 5 GiB) and setting KV_TARGET_GIB=16 to prevent memory watchdog failures.
  5. In comparisons with cloud models, the local model placed behind Kimi K3 (89.2%) but ahead of Qwen3.8-Max (79.6%) and GLM-5.2 (77.1%).

Why it matters

The run shows that high-grade local inference with a 262k context window can be hosted on a single GB10 unit without cloud API fees, provided the ~31 tokens/s decode speed fits the application profile.

Analysis

The benchmark underscores the trade-off between throughput and accuracy on single-node setups. While dual-node DeepSeek V4 setups offer faster decoding, Qwen3.8-Flash-Next provides stronger coding and reasoning outputs on less hardware. However, tight host memory tolerances mean practitioners must carefully calibrate KV cache targets and swap limits to prevent OOM failures.

Mentioned links

Found in

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • X post:Mike Gannotti

    Qwen3.8-Flash-Next on a Single NVIDIA DGX Spark Outscores Dual-Spark Setup in Benchmark

    A benchmark report from SMF Clearinghouse shows MiaAI's Qwen3.8-Flash-Next (NVFP4) scoring 87.3% on smf-bench Official A running on a single NVIDIA DGX Spark (GB10). The single-node serve outperformed the previous dual-Spark DeepSeek V4 Vision-Exp setup in overall accuracy, achieving a perfect score in coding.

    17bookmarks2.3kviews

    AI & LLMs· News

  • X post:Mia

    Performance Update for Qwen3.8 Flash Next on DGX Spark

    Mia (@MiaAI_lab) announced a major performance update for Qwen3.8 Flash Next running on a single DGX Spark, achieving up to 108 tok/s across concurrent streams and extending the KV cache to 1M.

    520bookmarks75.7kviews

    AI & LLMs· Announcement

  • X post:Mia

    Running DeepSeek v4 Flash on a Single NVIDIA DGX Spark

    MiaAI-Lab published an open-source launcher allowing DeepSeek V4 Flash 0731 (EXL3) to run on a single NVIDIA DGX Spark node instead of requiring a multi-node setup.

    71bookmarks9.8kviews

    AI & LLMs· Tool

  • Repository:MiaAI-Lab/DeepSeek-v4-Flash-One-DGX-Spark

    Running DeepSeek-v4-Flash on a Single DGX Spark

    An open-source launcher and Docker recipe for running DeepSeek V4 Flash 0731 (EXL3) on a single NVIDIA DGX Spark with 128 GiB unified memory.

    328starsPython

    AI & LLMs· Tool

  • Article:Elite Test Engineering

    Serving DeepSeek V4 Flash on Dual DGX Spark Nodes

    A detailed technical setup report documents serving DeepSeek-V4-Flash (284B MoE) across two DGX Spark nodes powered by Grace Blackwell chips using Ray and vLLM. It highlights workarounds for unified memory monitoring, NCCL RDMA deadlocks, and CUDA graph issues, as well as a 45% speedup using native speculative decoding.

    AI & LLMs· Tutorial

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.