Skip to main contentXiuAI/ XiuRouter
Unified AI model gateway

One API for Claude, GPT, and Gemini.

One interface, one ledger, one routing layer.The same first-party models, from subscription sources as low as 15% of list price—save on frequent calls, and switch back to the official provider in one click when a formal project needs a direct connection.

OpenAI-compatible, SDK-readyOne balance for all usage
Smart routing selects the more cost-effective source automatically

Choose by cost and stability without changing settings tool by tool

Original provider models

No replacement, no downgraded version, no watered-down model. Calls go to Claude, GPT, and Gemini themselves, not substitute models.

Usage can be checked call by call

Each call's model, tokens, and cost are written to the ledger, with transparent accounting for item-by-item reconciliation later.

Choose sources and switch as needed

By default, routing selects the best option based on cost and stability. For formal projects or specified procurement, switch to the official source in one click.

Pricing

Same first-party models, subscription sources as low as 15% of list.

Popular models from subscription sources go as low as 15% of list price. Figures below are public market reference prices (per million tokens); actual charges follow the per-call ledger in the console.

Popular models · per million tokens · USDView full price list
OpenAI · GPT-5.5
15% of list
openai/gpt-5.5
PriceSubscription sourceOfficial
Input$0.75/M$5.00/M
Output$4.50/M$30.00/M
Cache read$0.075/M$0.50/M
Anthropic · Sonnet 4.6
15% of list
anthropic/claude-sonnet-4.6
PriceSubscription sourceOfficial
Input$0.45/M$3.00/M
Output$2.25/M$15.00/M
Cache read$0.045/M$0.30/M
Google · Gemini 3.5 Flash
15% of list
google/gemini-3.5-flash
PriceSubscription sourceOfficial
Input$0.30/M$2.00/M
Output$1.20/M$8.00/M
Cache read$0.030/M$0.20/M

Get started in three steps

Sign up, add balance, and get an API Key.

1

Create an account

Create an account. Teams can create organizations for collaboration later.

2

Add balance

Add balance once and use it across all available models and sources.

3

Get an API Key

OpenAI-compatible. Paste it into your existing tools to start calling models.

Manage AI calls from one entry point.

Connect your everyday tools first, then choose a lower-cost or more stable path by task. Models, tokens, and costs all go into one ledger.