4× RTX 3090
87.5tok/sMixed
1 request, mixed prompt set, with Draft Source
Intelligenceno independent value
- Experimental
- License unclear
Local AI · Recipe · 1× DGX Spark
Laguna-S-2.1 EXL3 3.25 bpw with vLLM on 1× DGX Spark
by 0xSero
No speed value is documented for this recipe, so this page is not indexed.
Intelligence (original model) · Artificial Analysisno independent value
Engine
- Engine
- vLLM (EXL3-Tail-Overlay)
- Quantization
- Hybrid: 64 Hot-Experten NVFP4 + 192 Tail-Experten EXL3 Trellis-3 (nominal 3.25 bpw)
- Model family
- Laguna-S-2.1
- Context
- 200,000
- Creator
- 0xSero
- GitHub stars
- 6
- Repo updated
- Jul 26, 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 weights0xSero/Laguna-S-2.1-Hybrid-3.25bpw openmdw-1.1
- Basepoolside/Laguna-S-2.1-NVFP4 openmdw-1.1
- Engine
- vLLM (EXL3-Tail-Overlay)
- Context
- 200,000 tokens
Related recipes
1× DGX Spark
20.6tok/sMixed
1 request, mixed prompt set, with DFlash (per recipe) Source
Intelligenceno independent value
- Custom kernel required
- License unclear
4× RTX 3090
283tok/sPeak
1 request, code prompt Source
Intelligenceno independent value
- License unclear
1× RTX PRO 6000
155tok/sPeak
1 request, code prompt Source
Intelligenceno independent value
- License unclear
1× DGX Spark
76.7tok/sPeak
1 request, best value reported by source Source
Intelligenceno independent value
- License unclear
1× DGX Spark
53tok/sEveryday
1 request, prose prompt, with MTP (per recipe) Source
Intelligence40with thinking
- Custom kernel required
- License unclear
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.