4× RTX 3090
145tok/sEveryday
1 request, prose prompt, with DFlash (per recipe) Source
Intelligence34with thinkingno thinking 20
- Experimental
Local AI · Recipe · 2× RTX 3090
Qwen3.8 27B AutoRound INT4 with vLLM on 2× RTX 3090 (with DFlash)
by noonghunna
Intelligence (original model) · Artificial Analysis34with thinkingno thinking 20
Engine
- Engine
- vLLM
- Quantization
- AutoRound INT4 (W4A16) + fp8 KV + FA2-Plugin
- Model family
- Qwen3.8
- Context
- 262,144
- Parameters
- 27B dense
- Creator
- noonghunna
- GitHub stars
- 2,288
- Repo updated
- Sep 25, 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
- 2 × NVIDIA RTX 3090
- Weights
- Main weightsFrozenlock/Qwen3.8-27B-int4-AutoRound Apache-2.0
- Draft modelsyvai/Qwen3.8-27B-DFlash2-W4A16 Apache-2.0
- Engine
- vLLM
- Context
- 262,144 tokens
Related recipes
4× RTX 3090
85.3tok/sEveryday
1 request, prose prompt, with DFlash2 (per recipe) Source
Intelligence34with thinkingno thinking 20
2× RTX 3090
76.4tok/sEveryday
1 request, prose prompt, with MTP Source
Intelligence34with thinkingno thinking 20
1× RTX 3090
42.8tok/sEveryday
1 request, prose prompt, no speculative decoding Source
Intelligence34with thinkingno thinking 20
1× RTX 3090
39.7tok/sEveryday
1 request, prose prompt, no speculative decoding Source
Intelligence34with thinkingno thinking 20
1× DGX Spark
18.3tok/sEveryday
1 request, essay prompt, with DSpark Source
Intelligence34with thinkingno thinking 20
- 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.