Chat and coding models
Use GPT, Claude, Kimi, Qwen, DeepSeek, GLM and MiniMax models behind one consistent API surface.
- Chat, code and reasoning
- Provider-level routing
- Multi-key distribution
Aggregate many major AI model providers behind one OpenAI-compatible gateway, with user keys, PAYGO balance and usage accounting in the console.
Create an account first. No subscription required.
Register with an allowed email domain, then manage keys, balance and usage from the console.
Add USD-denominated credit through checkout and spend one balance across enabled model groups.
Use the OpenAI-compatible base URL in your app, CLI or agent workflow.
Use GPT, Claude, Kimi, Qwen, DeepSeek, GLM and MiniMax models behind one consistent API surface.
Test supported chat, coding and reasoning models from the console before moving traffic into production.

Point LangChain and agent frameworks at the same base URL while keeping balance, usage and routing centralized.
Send real prompts, attach supported files and inspect the experience before you wire the API into your product.
Top-up balance is spent across the model groups and routes your account can access.
User-facing keys are separated from upstream credentials, and sensitive values stay out of the browser.
No forced monthly plan. PAYGO balance is deducted from token and model usage.
Create separate keys for apps, experiments and production traffic.
High-volume users can contact us for routing, limits and pricing discussions.
Requests are distributed across configured upstream keys, with accounting kept at the user level.
CodexToken made our model access much easier to manage. One endpoint, one balance, and much less time spent switching configs.
The console is clean enough to test prompts quickly, then move the same setup into our agent workflow without extra glue code.
It is much easier to understand what our keys are spending. The PAYGO balance and usage view make day-to-day control simple.
CodexToken is PAYGO-first. You can create an account and open the console; model calls require available balance or granted credit.
No. PAYGO is the default flow, and an optional monthly plan is available when you want a separate subscription card.
Create separate keys for different projects. Rotate or disable a key if it appears in logs, clients or public repositories.
Usage is deducted from your user balance based on configured model pricing and token usage.
Yes. Contact us with your expected model mix, traffic pattern and reliability needs.
Send feedback through the console. New routes can be added when upstream availability and business rules allow it.
Use one console for user keys, balance, model access and usage tracking.
Create Account