Search Spins

DeepSeek debuts DeepSeek-V4.1-Flash, its smallest model built on a new Causal Encoder-Decoder architecture, with 552B backbone parameters and 1M-token context (Reuters)

DeepSeek launched DeepSeek-V4.1-Flash, a new small-language model built on an unverified 'Causal Encoder-Decoder' architecture, claiming 552B backbone parameters and 1M-token context length — positioning it as a compact yet high-capacity alternative in the open-weight LLM race.