Pay only for what you use. No hidden fees. No minimums.
For testing and small projects
For production workloads
Tailored for teams with specific needs
Tell us your use case, we'll craft a plan
Need more tokens? Buy a one-time top-up pack. No subscription, no expiration.
| Model | Input (1M tokens) | Output (1M tokens) |
|---|---|---|
| DeepSeek Chat (V3) | $0.27 | $1.10 |
| DeepSeek Reasoner | $0.55 | $2.19 |
| DeepSeek Coder | $0.27 | $1.10 |
| DeepSeek V4 Pro | $0.87 | $3.48 |
| DeepSeek V4 Flash | $0.27 | $1.10 |
| Qwen Max | $2.00 | $6.00 |
| Qwen Plus | $0.80 | $2.00 |
| Qwen3 235B-A22B | $1.20 | $3.50 |
| GLM-4 Plus | $1.00 | $5.00 |
| GLM-4 Air | $0.40 | $1.60 |
| GLM-4 Flash | Free* | Free* |
| Moonshot v1 (8K/32K/128K) | $1.20 / $1.80 / $3.00 | $6.00 / $9.00 / $15.00 |
| * GLM-4 Flash has free tier limits. Check provider for details. | ||
OpenRouter is great for model variety. ModelBridge is built for one thing: blazing-fast access to the best Chinese AI models, hosted on Hong Kong nodes with <100ms latency.
| Feature | ModelBridge | OpenRouter |
|---|---|---|
| Latency to Chinese models From overseas |
✅ <100ms Hong Kong node |
⚠️ 200–800ms Via US/EU nodes |
| Chinese model access DeepSeek, Qwen, GLM, Kimi |
✅ Primary focus | ✅ Available |
| Model catalog size | 12+ curated models | 300+ models |
| Pricing | Transparent per-token + sub | Provider price + 5.5% top-up fee |
| Subscription required? | ⚠️ Pro $9.90/mo (optional) | ✅ No subscription |
| Pay-as-you-go | ✅ Yes (no sub needed) | ✅ Yes |
| Support | ✅ Email + Dashboard | Community / Docs |
| Best for | Production apps using Chinese LLMs | Experimentation across many models |
Hong Kong deployment means <100ms to Chinese model APIs. OpenRouter routes through US/EU nodes — 3–8× slower.
We maintain direct connections to Chinese AI providers. No cooling periods, no surprise blocks.
Real-time token usage, cost breakdown, and billing history — built for production teams.
Email support on Pro plan. We help you debug integration issues, not just point to docs.
💡 Pro Tip: Start with our Free Tier — 500K tokens to test latency yourself. No credit card required.
Create Free AccountYou pay per token used. Monthly subscriptions include a token allowance — overage is billed at the per-token rates above. No hidden fees.
Yes! The free tier gives you 500K tokens/month at no cost. No credit card required.
We accept PayPal for all plans and token packs. Secure checkout with buyer protection included.
All connections use TLS 1.3 encryption. API keys are hashed in logs. Your prompts and responses are not stored beyond usage aggregation.
No! You can use ModelBridge without any subscription — just top up or pay per token. The $9.90/mo Pro plan is optional and adds quota + priority routing.