DeepSeek V4-Flash Released: AI Price War Intensifies

Chinese AI lab DeepSeek has released its official V4-Flash model, marking the latest development in an intense price and performance competition among open-source AI developers. This launch follows a preview of the V4 family in April. The V4 lineup includes two models. The V4-Pro is a larger model with 1.6 trillion total parameters and 49 billion active parameters. The V4-Flash is its leaner counterpart, featuring 284 billion total parameters with 13 billion active parameters. Both models offer a 1-million-token context window. On the Artificial Analysis Intelligence Index, V4-Pro scored 52. DeepSeek's own documentation notes that V4's capabilities are comparable to models released about six months earlier. What's interesting is that analysts see this release as strategically significant. V4-Flash offers enhanced agentic features and API pricing that is up to 50 percent cheaper than earlier versions. This move comes amid an accelerating price war among Chinese AI labs. The Council on Foreign Relations states that DeepSeek V4 signals a new phase in the US-China AI rivalry, driven by pricing strategy as much as model capability. This means more powerful AI could become much more accessible.