AI Model Price Competition

Back in late April, Xiaomi quietly released MiMo V2.5 Pro, a model that drew immediate comparisons to DeepSeek V4. Not a straight-up superior — more a lateral competitor, stronger in some areas, weaker in others — but priced noticeably lower. That asymmetry matters. You don't need to be better at everything when you're cheaper at everything. Then, by late May, the gap closed in the wrong direction for DeepSeek: MiMo 2.5 Pro's pricing had dropped to match DeepSeek V4 Pro's rates entirely. What started as a value proposition became a price war in the space of about a month.

If confirmed, here is what this means. Two capable models now sit at the same price point, which removes the only clean differentiator — cost — and forces developers and enterprises to make genuine capability trade-offs. That's a more honest market than the one we had six months ago, but it's a more brutal one too. DeepSeek built its reputation partly on being the disruptive underdog; having Xiaomi match it on price signals that the floor is moving fast. The second-order effect is pressure on everyone else. OpenAI, Anthropic, Google — each price cut in the Chinese market creates an awkward question about Western model pricing that doesn't get easier to answer.

Watch for any formal pricing announcements from either Xiaomi or DeepSeek that would confirm these figures, and pay close attention to whether other Chinese model providers follow suit in the next four to six weeks — that would signal a structural shift, not just a bilateral skirmish.

NewsHive monitors these sources continuously. All signal titles above link to the original reporting.

Intelligence by NewsHive. Need help navigating what this means for your business? Contact GeekyBee →