Deepseek makes 75% discount permanent, prices output 34x below GPT-5.5
deepseek permanently lowered v4 pro pricing to $0.87 per million output tokens, roughly 34.5 times cheaper than gpt-5.5.
deepseek announced that its v4 pro model discount is now permanent, offering input tokens for $0.435 per million and output tokens for $0.87 per million. compared to gpt-5.5, this represents roughly an 11.5x discount on input and a 34.5x discount on output tokens, with even greater savings against long-context pricing. although v4 pro lags behind frontier models like opus 4.7 in raw capability, the massive price gap makes it attractive for high-volume agentic applications. as roi becomes a priority, companies may increasingly shift from top-tier models to cost-effective alternatives, escalating the price war between chinese and western labs.