---
type: "Article"
title: "GLM-5.3-Flash EXL3 2.25 bpw with ExLlamaV3 on 1× DGX Spark (256k context)"
description: "GLM-5.3-Flash EXL3 2.25 bpw with ExLlamaV3 on 1× DGX Spark (256k context): recipe by r0b0tlab with speed, condition and source for every number, weights and license."
resource: "https://www.contextstudios.ai/local-ai/r0b0tlab--glm53-flash-exl3-2.25bpw-exllamav3-dflash2-1x-spark"
language: "en"
generated:
  by: "process:contextstudios-md/1"
  at: "2026-10-24T16:18:30.160Z"
status: "stable"
---

# GLM-5.3-Flash EXL3 2.25 bpw with ExLlamaV3 on 1× DGX Spark (256k context)

GLM-5.3-Flash EXL3 2.25 bpw with ExLlamaV3 on 1× DGX Spark (256k context): 20.6 tok/s according to github.com (dataset as of Sep 29, 2026).

- Creator: [r0b0tlab](https://www.contextstudios.ai/local-ai/creators/r0b0tlab.md)
- Engine: ExLlamaV3 v1.5.2 + TabbyAPI 816c321 (ghcr.io/r0b0tlab/glm53-flash-exl3-exllamav3-gb10:b25c467)
- Quantization: EXL3 2.25 bpw (Drafter EXL3 3.00 bpw), FP16 KV
- Model family: GLM-5.3-Flash
- Hardware: dgx-spark x1
- Context: 262144
- Artificial Analysis Intelligence Index: 41.8 ([source](https://artificialanalysis.ai/models/glm-5-3-flash))

## All numbers

- 20.6 tok/s (Everyday; Everyday, 1 request, prose prompt, with DFlash2 K=5 (per recipe)) — [source](https://github.com/r0b0tlab/glm53-flash-exl3-exllamav3-gb10/blob/main/docs/RESULTS.md#matched-8k-throughput-ar--dflash2-k5-median-of-5-c1)
- decode_tps: 20.56 (1 request) — [source](https://github.com/r0b0tlab/glm53-flash-exl3-exllamav3-gb10/blob/main/docs/RESULTS.md#matched-8k-throughput-ar--dflash2-k5-median-of-5-c1)
- decode_tps: 52.9 (1 request) — [source](https://github.com/r0b0tlab/glm53-flash-exl3-exllamav3-gb10/blob/main/docs/RESULTS.md#matched-8k-throughput-ar--dflash2-k5-median-of-5-c1)

## Weights

- [r0b0tlab/GLM-5.3-Flash-EXL3-2.25bpw-sm121](https://huggingface.co/r0b0tlab/GLM-5.3-Flash-EXL3-2.25bpw-sm121) (main, mit)
- [r0b0tlab/GLM-5.3-Flash-DFlash2-EXL3-3.00bpw](https://huggingface.co/r0b0tlab/GLM-5.3-Flash-DFlash2-EXL3-3.00bpw) (draft, other)
- [incoai/GLM-5.3-Flash-DFlash2](https://huggingface.co/incoai/GLM-5.3-Flash-DFlash2) (draft, cc-by-nc-nd-4.0)

[https://github.com/r0b0tlab/glm53-flash-exl3-exllamav3-gb10](https://github.com/r0b0tlab/glm53-flash-exl3-exllamav3-gb10)

## Related

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