DeepSeek releases V4.1-Flash, smallest of new architecture family with native vision

eliebakouch · x · 2026-09-10

DeepSeek released V4.1-Flash, the smallest model in its new architecture family, featuring native visual understanding. The company says the architecture is designed for greater capability, faster inference, higher throughput, and scaling to larger models. Researcher eliebakouch called it one of the most impressive architectures he has seen in a while.

Related event: DeepSeek Unveils Open-Source V4.1-Flash MoE Model(28 posts)→

Original post →

More from Models

Models channel →