4× RTX PRO 6000
235tok/sEveryday
1 request, prose prompt, context 19, with DSpark (per recipe) Source
Intelligence34with thinking
- Custom kernel required
Local AI · Recipe · 4× RTX PRO 6000
DeepSeek-V4-Flash FP8 with vLLM on 4× RTX PRO 6000
DeepSeek-V4-Flash FP8 with vLLM on 4× RTX PRO 6000: 215 tok/s according to github.com (dataset as of Nov 3, 2026).
by jcartu
Intelligence (original model) · Artificial Analysis24with thinkingno thinking 19
Engine
- Engine
- vLLM 0.21.1rc1.dev339+g1967a5627bc3
- Quantization
- FP8 W8A8 (nativ), FP8 KV block 256
- Model family
- DeepSeek-V4-Flash
- Context
- 1,048,576
- Creator
- jcartu
- GitHub stars
- 3
- Repo updated
- Jun 21, 2026
Measurements
Every sourced value of this recipe, each with its condition and source. Bars relative to the largest value in the group.
Decode · 1 request
Prefill by context
Time to first token
Shorter is better.
What you need
- Hardware
- 4 × NVIDIA RTX PRO 6000 Blackwell
- Weights
- Main weightsdeepseek-ai/DeepSeek-V4-Flash MIT
- Engine
- vLLM 0.21.1rc1.dev339+g1967a5627bc3
- Context
- 1,048,576 tokens
Sources
- github.com/jcartu/dsv4-flash-blackwell-inference-study#sustained-decode-cstechdevdsv4-flash-deepseek-v4-flash-fp8-tp4-mtp2
- github.com/jcartu/dsv4-flash-blackwell-inference-study#prefill-throughput
- github.com/jcartu/dsv4-flash-blackwell-inference-study
- github.com/jcartu/dsv4-flash-blackwell-inference-study/blob/main/data/benchmark_results_standard.json
Related recipes
4× DGX Spark
46tok/sEveryday
1 request, prose prompt, with DSpark Source
Intelligenceno independent value
4× DGX Spark
42tok/sEveryday
1 request, prose prompt, with DSpark (per recipe) Source
Intelligenceno independent value
- Custom kernel required
3× DGX Spark
41.9tok/sEveryday
1 request, prose prompt, with DSpark (per recipe) Source
Intelligenceno independent value
- Custom kernel required
2× DGX Spark
33.2tok/sEveryday
1 request, prose prompt, with DSpark (per recipe) Source
Intelligenceno independent value
- Custom kernel required
1× DGX Spark
30.2tok/sEveryday
1 request, story prompt, prompt 18 tokens, with MTP (per recipe) Source
Intelligence34with thinking
- ≤ 3 bit: quantization may cost quality
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.