io.net(@ionet):DeepSeek V4.1 Flash is live on https://t.co/DCp3kvjsTY. Day zero. Sparse MoE, 552B backbone, 8B/16B activation split, native image understanding, ~4x smaller KV cache than the previous Flash gen. New models drop. We ship the same day.

2026.09.10 21:04

DeepSeek V4.1 Flash is live on Day zero. Sparse MoE, 552B backbone, 8B/16B activation split, native image understanding, ~4x smaller KV cache than the previous Flash gen. New models drop. We ship the same day.

DeepSeek@deepseek_ai

2026.09.10 06:10

🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient. 🔹 Introducing the smallest model in our new architecture family, with native visual understanding. 🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models. 1/6