What This Is
On September 22, 2026, OpenAI and Anthropic released new products within roughly an hour of each other. OpenAI launched GPT-6 Sol and GPT-6 Luna, with API pricing cut in half versus the promotional price of the previous-generation GPT-5.6—and this is long-term pricing, not a limited-time offer. Anthropic released Claude Opus 5.5, which costs 40% less than Opus 5 to do the same work and runs 30% faster on output. Guangming Daily's next-day coverage made the point bluntly: both companies have made "cheaper" their core selling point this round.
But the editorial team is more interested in a different signal. Sam Altman framed this price cut as driving broader AI adoption and opening a new Renaissance. We notice an industry pattern: when a product has to halve its price to pull in users, it's usually not because the product got worse—it's because not enough people are buying. In other words, price is no longer the main barrier; the barrier has moved elsewhere. Where to? Our judgment: from "can afford it" to "can articulate it clearly."
Industry View
The bullish camp treats this as a pivotal moment for AI accessibility. Two major players cutting prices on the same day is being read as a shift from "competing on parameters" to "competing on penetration rate." Half-price means SMEs, indie developers, and one-person companies can more aggressively embed AI into workflows—the original article's author "Xiaohu" (小虎) is himself the case study, claiming he uses AI to work just three hours a day while maintaining daily output across the network.
But the dissenting voices are worth hearing too. The first objection: price cuts don't necessarily translate into deeper usage. Many people subscribe to ChatGPT for a year and end up only using it to chat, check the weather, and write email openings. The second, more sober view: model pricing was never the main obstacle to Chinese enterprise IT procurement—the obstacles are data compliance, private deployment, and employees not knowing how to use it. Cutting a bill in half just means a cheaper subscription fee for companies without clear AI use cases.
There's also an easily overlooked risk: the API price war squeezes model vendors' profit margins and may affect next-generation R&D investment over the long term. Smaller players who can't follow the price cuts may be forced out of the mainstream market, further raising industry concentration.
Impact on Regular People
For enterprise IT: Procurement budgets get short-term relief, but the real question is "do you have a fixed workflow on hand that AI can take over?" If you can't even produce that list, the saved API spend just makes idle subscriptions cheaper.
For individual careers: People who can articulate requirements will be more valuable than those who can use tools. The article redefines scarce ability from "knows how to use AI" to "can clearly explain when something starts, what gets delivered, how to know it's done right, and who signs off"—if you can't answer those four questions, even the strongest model is just guessing.
For the consumer market: The price war in AI subscription products will continue to trickle down—C-end users will see more promotions like "half-price memberships" and "doubled free quotas." But feature homogenization is accelerating, single-point tools are increasingly unable to build moats, and integrated workflow products may be the real winners of the next phase.