---
type: "Article"
title: "GLM-5.3 EXL3 3.38 bpw with TensorFold on 4× DGX Spark"
description: "GLM-5.3 EXL3 3.38 bpw with TensorFold on 4× DGX Spark: recipe by vcruz305 with speed, condition and source for every number, weights and license."
resource: "https://www.contextstudios.ai/local-ai/vcruz305--glm53-exl3-3.38bpw-tensorfold-tp4-4spark"
language: "en"
generated:
  by: "process:contextstudios-md/1"
  at: "2026-10-24T16:15:58.130Z"
status: "stable"
---

# GLM-5.3 EXL3 3.38 bpw with TensorFold on 4× DGX Spark

GLM-5.3 EXL3 3.38 bpw with TensorFold on 4× DGX Spark: 35.7 tok/s according to github.com (dataset as of Sep 29, 2026).

- Creator: [vcruz305](https://www.contextstudios.ai/local-ai/creators/vcruz305.md)
- Engine: TensorFold Fork vcruz305/TensorFold glm53-tp4-spark @757a851 (PR #159, TP=4) + b12x RoCE (b58f34e)
- Quantization: EXL3 SAGE MixedK 3.38 bpw (319 GB; serviert mit BF16-lm_head), bf16 KV, Draft DFlash2 d7/c0.60
- Model family: GLM-5.3
- Hardware: dgx-spark x4
- Context: 262144
- Artificial Analysis Intelligence Index: 44.8 ([source](https://artificialanalysis.ai/models/glm-5-3))

## All numbers

- 35.7 tok/s (Everyday; Everyday, 1 request, prose prompt, with DFlash2, thinking on (per recipe)) — [source](https://github.com/vcruz305/GLM-5.3-EXL3-DGX-Spark-recipe/blob/main/tensorfold-four-spark-tp4/README.md#decode-per-prompt)
- decode_tps: 35.71 (1 request) — [source](https://github.com/vcruz305/GLM-5.3-EXL3-DGX-Spark-recipe/blob/main/tensorfold-four-spark-tp4/README.md#decode-per-prompt)
- decode_tps: 41.11 (1 request) — [source](https://github.com/vcruz305/GLM-5.3-EXL3-DGX-Spark-recipe#performance)
- decode_tps: 56.5 (1 request) — [source](https://github.com/vcruz305/GLM-5.3-EXL3-DGX-Spark-recipe#performance)
- prefill_tps: 254.4 (1 request) — [source](https://github.com/vcruz305/GLM-5.3-EXL3-DGX-Spark-recipe#performance)

## Weights

- [vcruz305/GLM-5.3-EXL3-3.38bpw](https://huggingface.co/vcruz305/GLM-5.3-EXL3-3.38bpw) (main, other)
- [incoai/GLM-5.3-DFlash2](https://huggingface.co/incoai/GLM-5.3-DFlash2) (draft, cc-by-nc-nd-4.0)
- [zai-org/GLM-5.3](https://huggingface.co/zai-org/GLM-5.3) (base, other)

[https://github.com/vcruz305/GLM-5.3-EXL3-DGX-Spark-recipe](https://github.com/vcruz305/GLM-5.3-EXL3-DGX-Spark-recipe)

## Related

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