OpenAI-compatible API

One API for
Asia’s best AI models.

GLM, Kimi, DeepSeek and MiniMax on one key. It works like the OpenAI and Anthropic APIs, so you change one line and keep your code and tools.

Try for free

50% off every model, pay as you go. Up to 90% off for Enterprise. Start with $20 of free credit.

Top up from $10, or pay once for the Gold Pass and get $100 of usage a day in year 1.

Scroll to explore
5Accounts
5API Keys
5Bills
What is EastRouter

Why companies use EastRouter

One account instead of five: one key, one balance, one bill.

  1. 01

    One key and one bill for every model

  2. 02

    50% off every model, pay as you go

  3. 03

    New Asian models the day they launch

  4. 04

    Try any model on any task by changing one word

  5. 05

    One payment for years of use, with the Gold Pass

Models

The leading models from China and Asia.
One API.

GLM

GLM 5.2, GLM 5.3 Flash and GLM 5.3. GLM 5.3 Flash is the lowest-priced model on EastRouter, for jobs you run thousands of times. GLM 5.2 and 5.3 take the harder work.

Best for

GLM 5.2, GLM 5.3 Flash and GLM 5.3. GLM 5.3 Flash is the lowest-priced model on EastRouter, for jobs you run thousands of times. GLM 5.2 and 5.3 take the harder work.

Input from$0.075per 1M tokensOutput from$0.25per 1M tokens

Kimi

Kimi K2.6 and Kimi K2.7 Code. K2.6 writes articles, posts and emails. K2.7 Code is made for coding agents.

Best for

Kimi K2.6 and Kimi K2.7 Code. K2.6 writes articles, posts and emails. K2.7 Code is made for coding agents.

Input from$0.475per 1M tokensOutput from$2.00per 1M tokens

DeepSeek

DeepSeek V4 Pro and DeepSeek V4.1 Flash. Strong at reasoning, math and code, at a fraction of what the big Western models cost.

Best for

DeepSeek V4 Pro and DeepSeek V4.1 Flash. Strong at reasoning, math and code, at a fraction of what the big Western models cost.

Input from$0.15per 1M tokensOutput from$0.60per 1M tokens

Qwen

Multilingual work and coding. Tell us your workload and we’ll switch it on for your account.

Best for

Multilingual work and coding. Tell us your workload and we’ll switch it on for your account.

On requestAvailable upon request

MiniMax

MiniMax M3. For chat apps, agents and anything that talks to your customers.

Best for

MiniMax M3. For chat apps, agents and anything that talks to your customers.

Input from$0.30per 1M tokensOutput from$1.20per 1M tokens
Works with your stack

Change one line.
Keep your tools.

Works with the OpenAI and Anthropic SDKs, Claude Code, Codex, Cline, RooCode and your own agents. Point them at EastRouter and keep everything else.

View integration guides
Claude Code OpenAI SDK Codex CLI RooCode Cline Custom Agents

https://api.eastrouter.com/v1

Run your first task in seconds.

Get a key, point your tool at EastRouter and pick a model. No code to rewrite.

  1. Get an API key
  2. Point your tool at EastRouter
  3. Pick a model
Change just1line.
Try for free

$20 of free credit to try every model, for 14 days.

Full visibility

See what every request costs.

Each request is listed with its model, tokens in and out, latency, status and cost.

Recent requests

LiveSample data
TimeModelTokens in/outLatencyStatusCost
Customers

What changed after switching

Great for long-context work. We do repo-wide analysis and refactoring at a fraction of the cost.
Kevin ZhangIndie Developer
Switching to EastRouter was a no-brainer. Same coding quality as Claude, but my monthly bill dropped by 94%.
Alex MartinEngineering Manager
Common questions

What people ask first.

Can’t find what you need?

Read the docs
Which models can I access?

GLM, Kimi, DeepSeek and MiniMax, all on one key. Qwen on request. We add models as they launch; the models page has the full list.

What does it cost?

Pay as you go is 50% off the official rate on every model: every top-up is doubled, so $10 buys $20 of credit. Up to 90% off for Enterprise: email sales@eastrouter.com. Or pay once for the Gold Pass.

Do I need to change my code?

No. EastRouter works with the SDKs and tools you already use. Point them at EastRouter and keep your existing code.

How does the Gold Pass work?

You pay $4,999 once. The pass never expires and gives you a daily allowance on every model: $100 a day in year 1, $20 a day in years 2 to 6, then $5 a day. Nothing renews.

Can I compare and switch between models?

Yes. Change the model name in your request, and track latency and cost for every call in the dashboard.

How do I get started?

Sign up and you get $20 of free credit. Create a key, change the base URL in your tool and send your first request. It takes a few minutes.

Docs

Guides for every language you use

Quickstart, endpoint reference and setup for Python, JavaScript, TypeScript, Go, cURL and Rust.

Browse full documentation
Contact

Talk to our team.

Tell us what you’re building and we’ll reply by email.