DeepSeek Retires V4 Flash and V4 Flash Vision Exp, Redirects API Requests to V4.1 Flash

DeepSeek V4.1 Flash is now available on the DeepSeek API, while the legacy V4 Flash and V4 Flash Vision Exp models have been taken offline. Requests using the model names deepseek-v4-flash and deepseek-v4-flash-vision-exp will be temporarily routed to V4.1 Flash.

DeepSeek officially released V4.1 Flash on September 10, 2026, describing it as a 552B-parameter mixture-of-experts (MoE) model built on a new Causal-Encoder-Decoder architecture. It is the smallest model in DeepSeek's new architecture family. According to DeepSeek, V4.1 Flash outperformed flagships including DeepSeek V4 Pro, GLM5.3, and Kimi-K3 on Agentic Benchmark tests.

Compared with V4 Flash, V4.1 Flash reduces HBM demand to one-quarter and SSD demand to one-eighth. DeepSeek has also lowered V4.1 Flash prices. The new pricing took effect at 12:00 Beijing time on September 10, 2026: off-peak cached input is 0.02 yuan, uncached input is 1.0 yuan, and output is 4.0 yuan; peak-hour cached input is 0.04 yuan, uncached input is 2.0 yuan, and output is 8.0 yuan.

In addition, DeepSeek will phase out V4 Pro starting at 12:00 Beijing time on September 14, 2026. All deepseek-v4-pro requests will be routed to the new model and billed at V4.1 Flash rates.