Dev.to
6/24/2026

How to Use Chinese LLMs (Qwen, DeepSeek, GLM) Without a Chinese Phone Number
Short summary
Chinese LLMs like DeepSeek V4-Pro and Qwen 3.7-Max outperform GPT-4o on coding and reasoning while costing 80% less ($7/month vs $38), but require +86 Chinese phone numbers for registration. Gateway services solve this by maintaining Chinese infrastructure and handling billing, exposing everything through standard OpenAI-compatible APIs. Sign up via email/Stripe, swap your base_url, and immediately access better, cheaper models without VPN or complex workarounds.
- •Chinese models match or exceed GPT-4o performance while costing 80% less
- •Gateway services eliminate +86 phone requirement by hosting Chinese infrastructure
- •Switch to Chinese LLMs using a single base_url change in standard OpenAI SDK
Generated with AI, which can make mistakes.
Is this a good recommendation for you?



