One API.
All Models. Smart Routing.
China's top LLMs unified under a single endpoint. Write once, access all.
No more juggling API keys. No more rewriting integration code. FlintAPI aggregates DeepSeek, Qwen, Kimi, GLM, and MiniMax — and routes your requests to the optimal model automatically.
Why FlintAPI
Aggregate today. Route intelligently tomorrow.
Phase 1 — Model Aggregation
All major Chinese LLMs in one API endpoint. DeepSeek, Qwen, Kimi, GLM, MiniMax — switch models instantly. DeepSeek V4, Qwen3.5, Kimi K2, GLM-5, MiniMax M2 — switch models by changing one parameter. No library changes, no code rewrites.
- 🔌 OpenAI-compatible — drop-in replacement
- 💰 Competitive pricing — aggregated volume = lower cost
- 🔄 No vendor lock-in — switch models instantly
- 📊 Unified billing — one account, all models
Smart Routing
Just call /v1/smart/chat/completions or set model="flint-smart". FlintAPI auto-routes to the optimal model for your prompt. Call flint-smart for auto-routing. Or pick any model directly. analyzes your task — coding, translation, reasoning, creative writing — and automatically dispatches to the optimal LLM. Better results, lower cost.
- 🎯 Intent-aware routing — match task to model strength
- 💸 Cost-optimized — use cheap models for simple tasks
- ⚡ Latency-aware — fallback to faster models under load
- 📈 Quality-optimized — route hard problems to strongest model
All Major Chinese LLMs. One API.
No fragmentation. No multi-vendor complexity.
30 Seconds to Integrate
Fully OpenAI SDK compatible. Python, JavaScript, Go, cURL — pick your weapon.
from openai import OpenAI # One base_url. All models. client = OpenAI( base_url="https://www.flintapi.ai/v1", api_key="tk-your-api-key" ) # DeepSeek for code client.chat.completions.create( model="deepseek-v4-pro", messages=[{"role": "user", "content": "Write a Rust HTTP server"}] ) # Qwen for translation — same client, different model param client.chat.completions.create( model="qwen3.5-plus", messages=[{"role": "user", "content": "Translate to French: Hello World"}] )
model parameter to switch between any supported LLM. No SDK changes needed.
Simple, Transparent Pricing
Pay-as-you-go. No monthly minimums. Free credits on signup.
| Model | Context | Input /M tok | Output /M tok |
|---|---|---|---|
| DeepSeek V4 Pro | 128K | — | — |
| DeepSeek V4 Flash | 128K | — | — |
| Qwen3.5-Plus | 128K | — | — |
| Kimi K2.5 | 128K | — | — |
| GLM-5.1 | 32K | — | — |
Ready to unify your AI stack?
One API key. Five model families. Zero lock-in. Start building in 30 seconds.