Chinese-made AI models are becoming increasingly popular among U.S. companies as they narrow the performance gap with leading American competitors while remaining significantly cheaper to use.
Recent model releases from Chinese companies, including DeepSeek and Z.ai, are seen by many as highly competitive compared to leading frontier systems such as Anthropic and OpenAI. These performance advances come at a time when token prices for the most advanced models in many US AI labs are rising, leaving companies struggling with unexpectedly high costs associated with using the technology.
The share of tokens used by U.S. companies for Chinese AI models through OpenRouter – a platform that allows developers to access a range of AI models – has exceeded 30% every week since February 8, with that figure rising to 46%. The average over the last 12 months was just 11%, falling to 4.5% in the first half of 2025.
The rise of Chinese open source and open weight models comes as the U.S. government increasingly seeks to regulate its most powerful AI models and considers how to stop the rapid adoption of alternatives from overseas.
At the end of June, OpenAI announced that it would limit the introduction of new models at the request of the government. This month also saw the lifting of export controls on Anthropic’s Mythos and Fable models. after a tense standoff between the Trump administration and the company.
“Chinese AI models are particularly attractive to American companies now as AI costs are skyrocketing,” Kyle Chan, a fellow at the John L. Thornton China Center at think tank Brookings, told CNBC. “While U.S. companies previously prioritized AI adoption regardless of model, they are now becoming more cost-conscious.”
Table of Contents
Increasing acceptance
As companies seek to use AI models to develop new products and increase internal efficiency, engineers are increasingly experimenting with cheaper open-source and open-weight models, the most powerful of which are made by Chinese companies.
Open source and open weight models make different parts of an AI model available to developers to review, use, and sometimes modify. They differ from closed systems, like many flagship models from OpenAI, Anthropic and Googlewhere the code and inner workings remain proprietary.
In June, AI startup Lindy shifted 100% of its traffic from Anthropic’s Claude models to DeepSeek, a Chinese company that hit the market with a groundbreaking release in early 2025 and launched a new model in April.
“We did it, and you could see the cost curve going down, basically falling to the floor,” CEO Flo Crivello told CNBC. He said the decision would save Lindy millions of dollars within a few months.
OpenAI and Anthropic are facing a new AI reality as users move from “tokenmaxxing” to efficiency
DeepSeek saw its stake in gateway tokens increase on Vercel, a platform that enables developers to deploy and run apps and websites, between May and June.
Z.ai’s GLM 5.2 was released to great fanfare in June and saw the fastest rollout of any model tracked by Vercel in 2026, Harpreet Arora, head of agent infrastructure at Vercel, told CNBC. “In the first full week after launch, daily token volume grew approximately 27x and the number of customers using it increased approximately 80x.”
“Price is doing the work here,” Arora said. “If a task doesn’t require the best model, teams start to route them to the cheapest model that is good enough, and the recent wave of models from China are winning that trade.”
Chinese open source models can be “60 to 90% cheaper” than the leading Anthropic and OpenAI models, Justin Summerville, who works on data and analytics at OpenRouter, told CNBC.
OpenAI and Anthropic have been contacted for comment.
While Claude and ChatGPT still dominate in terms of usage on LaunchLemonade, an AI agent platform for regulated industries, GLM 5.2 is now among the top five models on the platform, Cien Solon, CEO and founder of LaunchLemonade, told CNBC.
“Chinese models like Z.ai and [Alibaba‘s] Qwen become options for companies [as] They offer an attractive combination of performance and cost for certain workloads,” Cien said. “Companies with more mature AI strategies are increasingly willing to deploy them where they make technical or economic sense.”
Approaching the border
The performance of Chinese AI models is also increasing.
While they are often “a fraction of the cost” of U.S. rivals, they operate “close to the leading American frontier models,” Brookings’ Chan said, estimating that they are currently “six to nine months” behind leading U.S. competitors.
“The new open source models perform well and prove suitable for all but the most complex LLM assignments,” Summerville said.
GLM 5.2 came in just a percentage point behind Anthropic’s Opus 4.8 in a closely watched agent benchmark and cost about a fifth of the price. Some researchers have said GLM 5.2 can compete with leading U.S. labs on some cyber benchmarks.
The move to DeepSeek V4 has increased performance in many core use cases for Lindy, Crivello said in a post on X.
“We’re seeing companies increasingly motivated to turn to cheaper AI stacks that they can control and customize themselves, and given the state of open source and open weight models, that often means leveraging Chinese options,” Yacine Jernite, head of machine learning at Hugging Face, told CNBC.
“There is a real risk that users will be stuck and have to choose between powerful but expensive proprietary US models, whose price and accessibility can fluctuate quickly, or that they will have to use Chinese models as the only possible alternative if they want to control costs or own their AI stack.”
Choose CNBC as your preferred source on Google and never miss a moment from the most trusted name in business news.
https://www.cnbc.com/2026/07/07/chinese-ai-models-costs-us-openai-anthropic.html
