Chinese AI company DeepSeek has released the DeepSeek-V4.1-Flash model with native multimodality and announced the phased retirement of V4-Pro. "Designed to expand capabilities, accelerate inference, increase throughput, and scale to larger models," the company stated.
The company described V4.1-Flash as the most compact model in its new architectural family. It is a Mixture of Experts neural network with 552 billion parameters and a Causal Encoder-Decoder architecture, with 8 billion parameters active on input and 16 billion on output. DeepSeek claims that through new pre-training methods and more extensive reinforcement post-training, the model has surpassed DeepSeek-V4-Pro in performance, cost, speed, and overall execution time.
DeepSeek also announced reduced KV cache requirements. V4.1-Flash requires 1/4 of the HBM volume and 1/8 of the SSD storage compared to the previous generation. The company believes this will reduce costs in agentic scenarios.
The model is already available via the DeepSeek API. DeepSeek has retired V4-Flash and V4-Flash-Vision-Exp — the compatible aliases deepseek-v4-flash and deepseek-v4-flash-vision-exp will temporarily redirect requests to V4.1-Flash. Starting at 04:00 UTC on September 14, all requests to deepseek-v4-pro will be automatically routed to V4.1-Flash at the new model's pricing rates. This mode will remain in effect until the launch of V4.1-Pro. The updated pricing for V4.1-Flash took effect at 04:00 UTC on September 10. DeepSeek also announced plans to work with the open-source community on supporting V4.1-Flash inference and to explore additional deployment options.
DeepSeek has hired CITIC Securities to prepare for an initial public offering on the STAR Market technology board of the Shanghai Stock Exchange, Reuters and SCMP report, citing their own sources. China's largest investment bank will serve as one of four underwriters. The company expects to begin the IPO process before the end of 2026. The exact timing of the listing, its size, and DeepSeek's valuation have not yet been determined. Chinese firms typically go through a preliminary preparation procedure involving brokerage structures before listing. The involvement of CITIC Securities indicates that DeepSeek has advanced further in its preparations for a potential offering.