OpenRouter, the aggregation platform that routes developers’ API calls across dozens of AI systems, reports that open-weight models created by Chinese labs now run more than three times the weekly code volume of US models.
The reason is that Chinese models that are roughly equivalent to their American counterparts cost a fraction. Price has become a deciding factor for developers optimizing with AI.
Why are Chinese AI models becoming more popular?
OpenAI and Anthropic sell access to closed models at a higher price per token, while labs like DeepSeek, Moonshot, and Zhipu publish weights that software developers can download and run without paying a per-token fee.
As of late June,… Six most used models on OpenRouter They were all open versions from Chinese companies, including Tencent, Xiaomi, DeepSeek, MiniMax, and Z.ai.
For context, in A Encrypted comparisonZhipu GLM 5.2 runs at $1.40 per million input tokens and $4.40 per million output tokens compared to $5 and $25 for Anthropic’s Opus 4.8.
During the week of February 9-15, 2026, Chinese models processed 4.12 trillion tokens on OpenRouter while US models handled 2.94 trillion. This was the first week that The American side of the industry lost. Through most of 2025, US systems accounted for nearly 70% of the top model’s usage.
Open-weight Chinese models also dominate downloads. Data shows it accounted for 41% of Hugging Face downloads this spring, outpacing US releases.
Vercel’s Production Index, published in July, found that open-weight models accounted for 29% of all tokens on its AI portal in June 2026, up from 11% in April. These models handled roughly a third of the tokens while withdrawing less than 4% of the spend. DeepSeek alone accounted for 22.6% of the gateway’s token volume, placing it in third place behind Google.
Why are industries adopting Chinese technology?
Cryptopolitan I mentioned A KPMG survey of 2,145 senior leaders found that 29% of them cannot understand or control the cost of operating their AI systems. Uber burned through it AI coding budget for 2026 By April, engineers were now capped at $1,500 per tool each month, while one unnamed company racked up a $500 million Cloud bill in a single month.
For example, Z.ai’s cloud API falls under China’s National Intelligence Law, meaning Chinese authorities can access any code or data sent through it. US lawmakers opened an investigation in May into risks posed by Chinese-origin models in critical infrastructure.
It is worth noting that Moonshot AI is preparing to launch Kimi K3, an open source version with 2.8 trillion parameters that the company says competes with the best American systems.
In terms of overall benchmarks, the model falls below Anthropic’s Claude Fable 5 and OpenAI’s GPT-5.6 Sol, but Cryptopolitan reported that it beat Claude Opus 4.8 and GPT-5.5 in coding and proxy tests. Kimi K3 is priced at $3,15 per million tokens, undercutting GPT-5.6 Sol at $5,30 and Fable 5 at approximately $10,50.
