2× DGX Spark
36.1tok/sEveryday
1 request, prose prompt, with DFlash2 (per recipe) Source
Intelligence42with thinking
- ≤ 3 bit: quantization may cost quality
- Custom kernel required
- Non-commercial
- License unclear
Local AI · Recipe · 4× RTX PRO 6000
GLM-5.3-Flash 320B-A18B NVFP4 with vLLM on 4× RTX PRO 6000
Engine
- Engine
- vLLM
- Quantization
- NVFP4 routed experts (QAD) + BF16 Rest, MXFP8 MTP-Experten; FP8 KV
- Model family
- GLM-5.3-Flash
- Context
- —
- Parameters
- 320B MoE / 18B active (lt. zai-org Model-Card)
- Creator
- local-inference-lab
- GitHub stars
- 3
- Repo updated
- Sep 19, 2026
Measurements
Every sourced value of this recipe, each with its condition and source. Bars relative to the largest value in the group.
No sourced measurements for this recipe.
What you need
- Hardware
- 4 × NVIDIA RTX PRO 6000 Blackwell
- Weights
- Main weightslocal-inference-lab/GLM-5.3-Flash-NVFP4 MIT
- Draft modellocal-inference-lab/GLM-5.3-Flash-DFlash2 Non-commercial · CC-BY-NC-ND-4.0
- Basezai-org/GLM-5.3-Flash-BF16 MIT
- Engine
- vLLM
Notes
What matters before you rebuild it.
- Experimental
- Overclocked memory
- Custom kernel required
- Non-commercial
Sources
- github.com/local-inference-lab/rtx6kpro/blob/master/models/glm-5.3-flash.md#measured-performance
- huggingface.co/local-inference-lab/lil-catalog/blob/main/GLM-5.3-Flash-NVFP4/lil.yaml
- github.com/local-inference-lab/rtx6kpro/blob/master/models/glm-5.3-flash.md
- huggingface.co/local-inference-lab/GLM-5.3-Flash-NVFP4
Related recipes
4× DGX Spark
34.1tok/sEveryday
1 request, prose prompt, with DFlash2 (per recipe) Source
Intelligence42with thinking
- Custom kernel required
- Non-commercial
4× DGX Spark
31.5tok/sEveryday
1 request, prose prompt, with DFlash2 (per recipe) Source
Intelligence42with thinking
- Custom kernel required
- Non-commercial
1× DGX Spark
19.5tok/sEveryday
1 request, prose prompt, with DFlash2 (per recipe) Source
Intelligence42with thinkingOriginal model's value. Heavily compressed (~2.3 bit), intelligence is likely much lower.
- ≤ 3 bit: quantization may cost quality
- Experimental
- Custom kernel required
- Non-commercial
2× DGX Spark
19.1tok/sEveryday
1 request, prose prompt, with DFlash2 (per recipe) Source
Intelligence42with thinking
- Non-commercial
- License unclear
2× RTX 3090
18.1tok/sEveryday
1 request, prose prompt, no speculative decoding Source
Intelligence42with thinking
- ≤ 3 bit: quantization may cost quality
- Partly in system RAM
Local AI in your company?
In a workshop we work out which models and which hardware fit your tasks, and build the first agent on your infrastructure.