DeepSeek V4 Flash (Variant 2)
DeepSeek / 1M
DeepSeek·1M context·Text
DeepSeek V4 Flash is a lightweight MoE reasoning model with 1M context, high throughput, automatic cache reads, and web search.
