DeepSeek's New AI Model Now Costs More Than Rival OpenAI's

DeepSeek's new V4 Flash model peaks at $1.32 per million words of output, surpassing OpenAI's budget GPT-5.6 Luna at $1.20 — a shift from DeepSeek's earlier strategy of charging far less than rivals. New rates begin August 16–17, 2026. The higher-tier V4 Pro rises from $0.87 to $3.96 per million tokens.