AI Wisdom

DeepSeek V4.1 Flash

The current DeepSeek API default — 552B MoE, 1M context, MIT license

Production
4/5
Open Source🧠 AI Engineering

Verdict

Replaced V4 Flash on the DeepSeek API on September 10, 2026; the old deepseek-v4-flash name still works but is served by V4.1 Flash, and the current model name is deepseek-flash. A 552B-parameter MoE with 8B active parameters for input and 16B for output, a 1M context, native image input, and MIT-licensed weights. Pricing is $0.30 input (cache miss) / $1.20 output per million tokens at peak hours and half that off-peak. DeepSeek says V4.1 Flash beats V4 Pro in its own and third-party tests, which makes it the default for high-volume pipelines.