Chinese AI Models Just Hit 46% of US Enterprise Tokens — Here’s Why Devs Are Ditching GPT-5.6
Author(s): Chew Loong Nian – AI ENGINEER
Originally published on Towards AI.
Chinese AI Models Just Hit 46% of US Enterprise Tokens — Here's Why Devs Are Ditching GPT-5.6
Chinese AI models peaked at 46% of US enterprise token usage in a single week this summer, according to a CNBC investigation of OpenRouter traffic published July 7. Eighteen months ago that number was 4.5%. The twelve-month average is 11%, and for every single week since February 8, 2026, Chinese-origin models have carried at least 30% of enterprise token volume on the largest neutral LLM router in the world.
After the initial lead, the article argues that the shift is driven less by hype or ideology and more by economics and operational fit: Chinese open-weight models consistently undercut leading US offerings (up to 60–90% cheaper, with output gaps reaching orders of magnitude). It backs this with real-world routing/usage examples, a detailed pricing comparison table, and a practical cost-calculation scenario for a production agent workload. The piece then addresses the main objection—security and government restrictions—by distinguishing between “Chinese model” and where models are served, emphasizing that open weights can be deployed on US-hosted or customer-owned infrastructure (e.g., Together/Fireworks/Bedrock or on-prem GPUs) to fit compliance needs. Finally, it provides a step-by-step 5-minute evaluation method: A/B test your own prompts via OpenRouter, route with fallbacks (LiteLLM) to preserve flagship quality for hard cases, and adjust deployment options to satisfy compliance constraints, concluding with a verdict that the open-weight price advantage is quickly outweighing small benchmark differences for most enterprise tasks.
Read the full blog for free on Medium.
Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.
Published via Towards AI
Towards AI Academy
We Build Enterprise-Grade AI. We'll Teach You to Master It Too.
15 engineers. 100,000+ students. Towards AI Academy teaches what actually survives production.
Start free — no commitment:
→ 6-Day Agentic AI Engineering Email Guide — one practical lesson per day
→ Agents Architecture Cheatsheet — 3 years of architecture decisions in 6 pages
Our courses:
→ AI Engineering Certification — 90+ lessons from project selection to deployed product. The most comprehensive practical LLM course out there.
→ Agent Engineering Course — Hands on with production agent architectures, memory, routing, and eval frameworks — built from real enterprise engagements.
→ AI for Work — Understand, evaluate, and apply AI for complex work tasks.
Note: Article content contains the views of the contributing authors and not Towards AI.
Related posts
Recent Posts
