If you're like most developers, you probably started with OpenAI's API. It's the default choice — great documentation, solid models. But as your project grows, you might notice some friction:
- Costs add up fast. A single chat-heavy app can burn through hundreds of dollars.
- No fallback. If OpenAI goes down, your app goes down.
- One model doesn't fit all. Sometimes you need a fast model for simple queries and a reasoning model for complex tasks.
Powered by some of the most advanced AI models from China's leading labs — giving you enterprise-grade performance at a fraction of the usual cost.
The solution? A multi-model API gateway that gives you access to multiple providers through a single OpenAI-compatible endpoint.
What You'll Need
- An API key from Dubhe Hub (free tier includes 100K tokens)
- Any OpenAI SDK (Python, Node.js, curl — they all work)
Step 1: Get Your API Key
Go to dubhehub.com, sign up, and create an API key. The dashboard shows you six models to choose from.
Step 2: Make Your First Call
Here's a Node.js example. Notice that the code looks exactly like OpenAI's SDK:
import OpenAI from 'openai';
const dubhe = new OpenAI({
baseURL: 'https://dubhehub.com/v1',
apiKey: 'your-key-here'
});
// Fast model for quick responses
const quick = await dubhe.chat.completions.create({
model: 'dubhe-fast',
messages: [{ role: 'user', content: 'Explain REST APIs in 3 bullet points' }]
});
console.log(quick.choices[0].message.content);
Step 3: Swap Models Without Changing Code
The real power is switching models by changing just the model name:
// Use dubhe-reasoning for complex logic
const analysis = await dubhe.chat.completions.create({
model: 'dubhe-reasoning',
messages: [{ role: 'user', content: 'Write a debounce function in TypeScript' }]
});
// Use dubhe-vision for image analysis
const vision = await dubhe.chat.completions.create({
model: 'dubhe-vision',
messages: [
{ role: 'user', content: [
{ type: 'text', text: 'What's in this image?' },
{ type: 'image_url', image_url: { url: 'https://example.com/photo.jpg' } }
]}
]
});
Model Overview
| Model | Best For | Input/1M tokens | Output/1M tokens | Context |
|---|---|---|---|---|
| dubhe-fast | General chat, text | $0.30 | $0.60 | 1M |
| dubhe-code | Programming tasks | $0.80 | $3.00 | 200K |
| dubhe-agent | Automation, agents | $1.00 | $4.00 | 205K |
| dubhe-plus | Long context, analysis | $2.00 | $8.00 | 1M |
| dubhe-vision | Image understanding | $5.00 | $15.00 | 1M |
| dubhe-reasoning | Deep reasoning | $6.00 | $18.00 | 1M |
Why Multi-Model?
- Cost optimization: Use cheap models for simple tasks, expensive ones only when needed
- Provider diversity: If one provider has an outage, switch to another instantly
- Performance matching: Match model capability to task complexity
The migration takes about 5 minutes — just change the baseURL and apiKey. Everything else (streaming, function calling, embeddings) works the same way.
Try it yourself at dubhehub.com.
Top comments (0)