Provider pricing

SiliconFlow AI API pricing

SiliconFlow is a Chinese serverless inference provider running 100+ open-source models (Llama, Qwen, DeepSeek, GLM, Kimi, FLUX, Stable Diffusion) at some of the lowest per-token rates in the market. The OpenAI-compatible API includes function calling, JSON mode, and a generous free tier.

Founded2023
HeadquartersBeijing, China
Tracked modelsDirectory

Best for

Cost-sensitive teams and Asia-Pacific apps running open-source inference at scale.

Strengths

  • Among the lowest per-token rates on Llama, Qwen, and DeepSeek
  • Free tier with high rate limits for prototyping
  • OpenAI-compatible API with function calling and JSON mode
  • Strong coverage of Chinese open-weight models (Qwen, GLM, Kimi)