One endpoint, many models
A single OpenAI-compatible API. Switch between GLM, Kimi, Qwen and DeepSeek coding models without rewriting your code.
Official channels · Full power · OpenAI & Anthropic compatible
An
LLM API platform
Call DeepSeek, GLM, Kimi and Qwen coding models through official channels at full power, with no downgrades. Works with OpenAI- and Anthropic-compatible clients, priced lower than official channels with transparent per-token billing, and a drop-in setup built for coding and code tasks.
Works with the OpenAI SDK and Claude Code · swap the Base URL and key
Supported models
GLMDeepSeekKimiQwen
Setup environment…
# macOS / Linux
$ export ANTHROPIC_BASE_URL=https://api.coderoute.cc
$ export ANTHROPIC_AUTH_TOKEN=sk-xxx
# Windows PowerShell
PS> $env:ANTHROPIC_BASE_URL = "https://api.coderoute.cc"
PS> $env:ANTHROPIC_AUTH_TOKEN = "sk-xxx"
# Get started with Claude Code
$ claude
✨ Ready to build amazing things! Supported models 
Leading coding models, one endpoint
Built for coding, with a focused lineup of mainstream models. Available versions and pricing are listed in the dashboard.
GLM
Zhipu AI
Zhipu AI official models — strong coding and agent capabilities, with long-context support.
View pricing → DeepSeek
DeepSeek
DeepSeek open-source models — excel in math and code, with hybrid reasoning.
Open dashboard → Kimi
Moonshot AI
Moonshot AI long-context models, built for long-horizon coding and agent tasks.
Open dashboard → Qwen
Alibaba Cloud
Alibaba's Qwen series — open-source, covering code, vision and multimodal.
Open dashboard → Why CodeRoute 01 02 03 04 05 06
An LLM API platform that keeps coding simple
Full-power calls through official channels. No sprawling model mall, just coding done right: multi-model switching, OpenAI compatibility, transparent pricing.
Official channels, full power
Traffic goes to the official model channels, not reverse-engineered or downgraded endpoints. You get the same capability as the source model.
Transparent pricing
Pay per token, with clear per-model pricing you can check in the dashboard.
Drop-in setup
OpenAI-SDK compatible. Change your base URL and API key and you are done.
Smart routing
Multi-account scheduling and load balancing reduce the impact of a single endpoint going down.
Built for developers
Clear usage stats and API key management, well suited to solo developers and small teams.
Pricing Recommended 
Recommended 
Transparent pricing, cheaper than official
Billed per token — pay for what you use, cheaper than official channels. Works with Claude Code and the OpenAI SDK out of the box.
Pay-as-you-go
Never expires
No subscription — pay for actual usage
- Top up to get equivalent token credit
- Pay for actual usage
- Balance never expires
Claude Code pay-as-you-go
1:1 (RMB:USD)
No subscription — pay for actual usage
- 1 RMB = 1 USD, official price parity
- Supports mainstream coding models
- Optimized for Claude Code
OpenAI SDK pay-as-you-go
1:1 (RMB:USD)
No subscription — pay for actual usage
- 1 RMB = 1 USD, official price parity
- OpenAI-compatible protocol with all coding models
- Optimized for OpenAI SDK, Cursor and Codex
Three steps
From signup to your first call
Sign up and create an API key
Register in the dashboard and generate an API key in one click.
Set environment variables
Point ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN at CodeRoute.
Make the call
Send a coding request with supported models.
Start coding with an LLM API
Create an account and make your first call in minutes.
FAQ
Frequently asked questions
Common questions about the LLM API and integrating CodeRoute.
What is an LLM API?
An LLM API lets your application call a large language model over an interface. You send a request with an API key and get back text generation, code generation or reasoning results, usually billed per token used.
Which models does CodeRoute offer?
CodeRoute focuses on coding and provides mainstream coding model APIs. The exact list of available models and versions is shown in the dashboard.
Is CodeRoute OpenAI-compatible?
Yes. CodeRoute supports both the OpenAI-compatible and Anthropic-compatible protocols (ANTHROPIC_BASE_URL), so mainstream AI coding tools like Claude Code, VS Code, opencode, Codex, OpenClaw and Hermes connect directly — just swap the base URL and API key.
How is the LLM API billed?
You pay per token, and each model is priced differently. Per-model pricing is visible in the dashboard. There are no hidden fees.
How do I get started?
Create an account → generate an API key in the dashboard → point your base URL at CodeRoute and call with your key. Three steps and you are integrated.
Are the models the full official versions?
Yes. Every model on CodeRoute runs through official channels that connect directly to the model provider, never reverse-engineered or downgraded. Capability matches the original.
How does CodeRoute compare to official pricing?
CodeRoute is cheaper than the official channels. Discounts vary by model — see the dashboard for real-time per-model pricing. Billed per token, no hidden fees.