New Kimi K2 Thinking & Qwen3-Max are live. Browse models →

BrushLLM

About

China’s AI, served at thesource— for everyone

DeepSeek, Qwen, GLM, Kimi — the labs behind these models are pushing the frontier of coding, reasoning and agentic AI. But if you’re a developer outside mainland China, actuallyusing them means juggling local accounts, payments and docs in a language your stack doesn’t speak.

BrushLLM exists to close that gap. We aggregate 24 models from10 leading Chinese AI labs behind a single OpenAI-compatible endpoint, billed in USD, documented in English, and reachable from anywhere. Your code stays the same — only the base URL changes.

Our mission is simple: make the source of Chinese AI affordable and effortless for the world — every country, every region, every developer with an idea worth building.

Source for the world

China’s frontier models shouldn’t need a local ID, a local card, or a VPN to reach — and shouldn’t cost a fortune either. We deliver them, at the source, at prices any developer can afford.

Radical simplicity

One key, one endpoint, one bill. If integrating us takes longer than a coffee break, that’s our bug, not your problem.

Honest pricing

Published per-token rates, no subscription traps, and free tiers that stay free. You should always know what a request costs.

Talk to us

Questions, volume pricing, or a model you’d like us to add — we read everything.