DeepSeek Returns With V4 Pro and Flash, Closing the Gap on Frontier AI Models
In Brief: DeepSeek has released preview weights for V4-Pro (1.6T parameters, 49B active) and V4-Flash (284B / 13B active) under the MIT licence — topping all open-source models on coding and maths benchmarks, and self-reported within "3 to 6 months" of GPT-5.4 and Gemini 3.1-Pro.
China'