Alibaba hikes DeepSeek V4 Flash 0731 output prices by 230%

Alibaba sharply increased prices for DeepSeek V4 Flash 0731 overnight, raising output costs by 230% (from $0.40 to $1.32 per million tokens) and input costs by 120%, with similar hikes across other providers like Mistral and Qwen. The price increases signal the end of subsidized AI inference costs and force builders relying on cheap API access to reconsider their business models and optimize token usage. This shift reflects AI labs realigning pricing to reflect actual inference costs rather than pursuing a race-to-the-bottom pricing strategy.