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 · 1× DGX Spark
GLM-5.3-Flash Q2_K with llama.cpp on 1× DGX Spark
by vcruz305
17.6tok/sEveryday
1 request, realistic prompt, context 2k, with DFlash2 (per recipe)
Source: github.comIntelligence (original model) · Artificial Analysis42with thinking
≤ 3 bit: quantization may cost quality
Engine
- Engine
- llama.cpp Fork 6f5ac9a + 4a06ec6
- Quantization
- GGUF Q2_K (2.91 BPW) + Draft Q4_K_M/BF16, q8 KV
- Model family
- GLM-5.3-Flash
- Context
- 98,304
- Creator
- vcruz305
- GitHub stars
- 7
- Repo updated
- Sep 13, 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
- 1 × NVIDIA DGX Spark (GB10)
- Weights
- Main weightsvcruz305/GLM-5.3-Flash-GGUF MIT
- Draft modelvcruz305/GLM-5.3-Flash-DFlash2-GGUF Non-commercial · CC-BY-NC-ND-4.0
- Draft modelincoai/GLM-5.3-Flash-DFlash2 Non-commercial · CC-BY-NC-ND-4.0
- Visionvcruz305/GLM-5.3-Flash-GGUF MIT
- Basezai-org/GLM-5.3-Flash MIT
- Engine
- llama.cpp Fork 6f5ac9a + 4a06ec6
- Context
- 98,304 tokens
Notes
What matters before you rebuild it.
- ≤ 3 bit: quantization may cost qualityAt 3 bit and below the model may answer noticeably worse than the original. The intelligence number refers to the original.
- Non-commercial (draft)
- Custom kernel required
- Non-commercial
Sources
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.