DeepSeek-V4.1-Flash performance has been significantly improved.

1 reports1 outlets3 days agoUpdate

The story so far

AI overview

DeepSeek-V4.1-Flash achieves a service capability of 1.7M tokens/s per GPU in a single NVIDIA Vera Rubin NVL72 rack, reaching up to 387 tokens/s per user, representing a 58-fold improvement over GB300 NVL72. This is the result of DeepSeek optimizing models and hardware combined, aiming to enhance the efficiency of large model inference.

AI-written from the reports · Updated 2 days ago

Latest12 Nov, 08:00

DeepSeek-V4.1-Flash achieved a 58-fold throughput improvement on NVIDIA Vera Rubin NVL72.

Coverage timeline

Follow the coverage to see every side of the story.

12 Nov

08:00
vLLM Blog
DeepSeek-V4.1-Flash achieved a 58-fold throughput improvement on NVIDIA Vera Rubin NVL72.
DeepSeek-V4.1-Flash achieves a service capability of 1.7M tokens/s per GPU in a single NVIDIA Vera Rubin NVL72 rack, reaching up to 387 tokens/s per user, an increase of 58 times compared to GB300 NVL72.

Heat over time

Now: 3
Peak: 513 Nov, 07:00
24h change: -49%

Hover or tap the chart to see each hour; use the arrow keys on a keyboard.