DeepSeek permanently slashed prices for V4-Pro, charging just $0.435 per million input tokens. This pricing sits at least 34 times below GPT-5.5 for output tokens. The aggressive move targets high-volume agentic systems. Such deep cuts force Western providers to either lower their margins or find new ways to justify premium pricing.