DeepSeek V4.1 Flash

With a 1M context and 384K of output, DeepSeek V4.1 Flash now absorbs traffic once handled by V4 Pro.

DeepSeek put the first landing point of its new architecture on the smallest model. V4.1 Flash handles native vision, a 1M context and 384K of output, with API concurrency of 2500. The company says it beats V4 Pro on performance, cost, speed and total time, and from September 14 V4 Pro requests route to it. The line proves native vision and throughput on a small model first.

DeepSeek officially released V4.1 Flash: Native Vision, with the new architecture starting from small models

DeepSeek officially released DeepSeek-V4.1-Flash . Officially defined as the smallest model in the new architecture series, natively supporting multim...

Admin 9/10/2026

DeepSeek V4 Flash Vision opens weights: multimodal agents move toward on-premises deployment

On August 31, 2026, DeepSeek launched DeepSeek-V4-Flash-Vision-Exp open weights on the official Hugging Face organization deepseek-ai. The model repos...

Admin 9/1/2026

Latest AI news: DeepSeek launches V4-Flash-Vision-Exp, bringing multimodal vision capabilities into Flash models

In the past 24 hours (August 21 to August 22, 2026), the AI industry continued to compete around multimodal models, agents, cybersecurity, and computi...

Admin 8/22/2026

149

DeepSeek opens DSpark acceleration components: the model hasn't changed, but why does the generation speed improve?

On June 28, 2026, the official DeepSpec repository for DeepSeek was updated and added to the DSpark checkpoint, providing speculative decoding support...

Admin 6/28/2026

202

DeepSeek V4 is officially released: 1M context and open source weight are in focus

On April 24, 2026, DeepSeek V4 has officially surfaced. Unlike a few hours ago, when it could only be judged based on media rumors, now DeepSeek's off...

Admin 4/24/2026

660