Open AI Models Overtake Closed Ones: The Real Race May Not Be at the Frontier
Chinese open-weight AI models accounted for 41% of Hugging Face downloads this spring, surpassing US models. Companies increasingly prefer cheaper, customizable open models over expensive proprietary ones, raising questions about the future of frontier models.
While the industry focused on Anthropic's latest frontier models and Washington's access control debate, developers kept building with open solutions. This spring, Chinese open-weight models made up 41% of all downloads on Hugging Face, exceeding US models. On OpenRouter, the top six most popular models are open models from Chinese firms like Tencent, Xiaomi, DeepSeek, MiniMax, and Z.ai, with Anthropic's Claude Opus 4.7 trailing in seventh place.
Data from Vercel shows open models handled nearly a third of AI requests in June. Although these platforms exclude sessions from major labs (which likely account for most OpenAI and Anthropic usage), the growing market share of open models raises a critical question: how relevant are frontier models if most production AI runs on cheaper, customizable alternatives?
Hugging Face CEO Clem Delangue notes that companies increasingly prefer owning their AI models rather than renting them. A new repository is created every seven seconds on Hugging Face, which hosts nearly three million public models and one million public datasets. Half of Fortune 500 companies use Hugging Face to deploy private or open-source models.
Chinese AI labs are regularly releasing powerful open-weight models that are cheaper and easier to customize than closed competitors. Most recently, Beijing-based Z.ai released GLM-5.2, which excels at agentic coding and competes with Anthropic's models in identifying security vulnerabilities.
Microsoft CEO Satya Nadella warned against single provider lock-in, arguing that control of data should be a primary concern. He advocated for distributing learning infrastructure so every firm controls its own learning loop.
Debate continues over whether powerful open models should be widely available. Anthropic CEO Dario Amodei argues they could become dangerous if released, while Delangue contends that the biggest risk is power concentration. He believes transparency makes the world safer by allowing defenders to patch cybersecurity risks more easily.
