Chinese AI models regularly pass 30 percent on OpenRouter as cost gap widens

2026-07-07 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Fundamental Awareness, quick

Summary

Chinese AI models are increasingly adopted by US companies due to significantly lower costs compared to systems from OpenAI and Anthropic, as reported by CNBC. On the OpenRouter platform, Chinese models from providers like DeepSeek and Z.ai have consistently captured over 30 percent of traffic weekly since February 8, occasionally reaching 46 percent. This marks a substantial increase from last year's average of just 11 percent. OpenRouter employee Justin Summerville notes that Chinese open-source models are 60 to 90 percent cheaper. For instance, the startup Lindy fully transitioned its traffic from Anthropic's Claude to DeepSeek, with CEO Flo Crivello stating this switch saves millions. While Chinese models are competitive on cost, the Brookings Institution and the Center for AI Standards and Innovation (CAISI) estimate they trail leading US models by six to nine months in capabilities across areas like cybersecurity, software development, and abstract reasoning.

Key takeaway

For Directors of AI/ML evaluating model providers, the significant cost advantage of Chinese AI models like DeepSeek presents a compelling alternative. Despite a reported 6-9 month capability lag behind US leaders, their 60-90 percent lower per-token pricing can yield millions in savings, as demonstrated by Lindy's switch. You should assess these models for non-critical or cost-sensitive workloads to optimize your operational budget and potentially accelerate deployment.

Key insights

Chinese AI models are rapidly gaining US market share due to their 60-90% lower cost, despite a 6-9 month capability gap.

Principles

• Cost-efficiency drives AI model adoption.
• Capability gaps can be offset by pricing.
• Open-source models offer significant savings.

In practice

• Evaluate Chinese models for cost savings.
• Monitor OpenRouter for traffic trends.
• Consider DeepSeek for production workloads.

Topics

• Chinese AI Models
• AI Model Costs
• OpenRouter
• DeepSeek
• AI Market Share
• Model Performance Benchmarks

Best for: CTO, Investor, Entrepreneur, Director of AI/ML, VP of Engineering/Data, Consultant

Related on AIssential

• When Americans choose Chinese AI — Cost-effective Chinese AI models are gaining traction among U.S. developers, challenging established American providers despite geopolitical concerns.
• China is falling behind in the AI race, according to a US government benchmark — Chinese AI models lag US counterparts in capability but offer significant price advantages.
• not much happened today — AI development is shifting towards multi-cloud distribution, cost-aware agentic workflows, and specialized hardware for inference.
• The mysterious Hy3 LLM is topping OpenRouter Model Rankings by a large margin — Effective LLM pricing, driven by prompt caching and high input token usage, significantly deviates from stated costs, influencing model adoption.
• DeepSeek STRIKES BACK - Open-Source Model Challenges OpenAI And Anthropic — DeepSeek V4 sets a 1 million token context as standard, democratizing advanced AI capabilities through open-weight MoE models.
• China Takes Over — Deepseek V4 offers frontier-level AI performance at a fraction of the cost, challenging established market leaders.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.