DeepSeek's V4 Flash Now Costs More Than OpenAI's Budget Model

DeepSeek's V4 Flash API peaks at $1.32 per million output tokens, topping OpenAI's budget-tier GPT-5.6 Luna at $1.20 — a reversal from DeepSeek's earlier aggressive undercutting. Peak and off-peak rates take effect August 16–17, 2026. V4 Pro peak pricing rises from $0.87 to $3.96 per million tokens.