---
type: "Article"
title: "GLM-5.3-Flash NVFP4 with SparkRing on 2× DGX Spark"
description: "GLM-5.3-Flash NVFP4 with SparkRing on 2× DGX Spark: recipe by fujitsupolycom with speed, condition and source for every number, weights and license."
resource: "https://www.contextstudios.ai/local-ai/fujitsupolycom--glm-5.3-flash-nvfp4-tp2-2x-dgx-spark"
language: "en"
generated:
  by: "process:contextstudios-md/1"
  at: "2026-10-24T16:14:32.007Z"
status: "stable"
---

# GLM-5.3-Flash NVFP4 with SparkRing on 2× DGX Spark

GLM-5.3-Flash NVFP4 with SparkRing on 2× DGX Spark: 36 tok/s according to github.com (dataset as of Sep 29, 2026).

- Creator: [fujitsupolycom](https://www.contextstudios.ai/local-ai/creators/fujitsupolycom.md)
- Engine: vLLM (local-inference-lab Fork) via SparkRing (switchless NCCL ueber ConnectX-7)
- Quantization: NVFP4 Spark (Local Inference Lab)
- Model family: GLM-5.3-Flash
- Hardware: dgx-spark x2
- Artificial Analysis Intelligence Index: 41.8 ([source](https://artificialanalysis.ai/models/glm-5-3-flash))

## All numbers

- 36 tok/s (Everyday; Everyday, 1 request, decode at 16K context (profile table), thinking always, max; peak C=8 (profile serves max 8 requests)) — [source](https://github.com/FujitsuPolycom/sparkring#profiles)
- decode_tps: 36 (1 request) — [source](https://github.com/FujitsuPolycom/sparkring#profiles)
- decode_tps: 100 (16 parallel requests) — [source](https://github.com/FujitsuPolycom/sparkring#profiles)
- prefill_tps: 2460 (1 request) — [source](https://github.com/FujitsuPolycom/sparkring#profiles)

## Weights

- [local-inference-lab/GLM-5.3-Flash-NVFP4-Spark](https://huggingface.co/local-inference-lab/GLM-5.3-Flash-NVFP4-Spark) (main, mit)

[https://github.com/FujitsuPolycom/sparkring](https://github.com/FujitsuPolycom/sparkring)

## Related

- [Local AI on your hardware.](https://www.contextstudios.ai/local-ai.md)