Claude
- claude-fable-5-1Save 77%
- claude-fable-5Save 77%
- claude-opus-5Save 77%
- claude-sonnet-5Save 77%
XiuRouter · model API access
Claude, GPT and more through one API
Lower model costs, pay as you go. Follow the setup guide for your coding tool and check the cost of each request.
Reference prices, current rates, and savings
Enter your monthly input budget at the provider-reference rate to estimate XiuRouter input costs.
Price comparison
USD per 1M tokens
| Model | Reference input | Tier input | Input savings | Reference output | Tier output | Output savings |
|---|---|---|---|---|---|---|
| claude-fable-5-1 | $10.00 | $2.30 | Save 77% | $50.00 | $11.50 | Save 77% |
| gpt-6-astra | $10.00 | $0.88 | Save 91.3% | $50.00 | $5.25 | Save 89.5% |
| deepseek-v4-pro | $0.66 | $0.39 | Save 40.7% | $1.98 | $0.78 | Save 60.5% |
| claude-fable-5 | $10.00 | $2.30 | Save 77% | $50.00 | $11.50 | Save 77% |
| claude-opus-5 | $5.00 | $1.15 | Save 77% | $25.00 | $5.75 | Save 77% |
Open the pricing page to confirm current models and rates.
Savings vary by model and price type; check the current pricing table for each model.
Popular models and current pricing
Browse available models and savings by provider.
OpenAI
- gpt-6-astraSave 91.3%
- gpt-5.6-solSave 91.3%
- gpt-5.6-terraSave 93%
- gpt-5.6-lunaSave 93%
Most used in the last 7 days
- 01gpt-5.6-sol41.9%
- 02gpt-6-astra29.9%
- 03claude-opus-57.3%
Test, monitor, and compare
Test a request in the Playground, inspect costs in Usage, and see popular models in Rankings.
01
Test a request
Choose a model and inspect a response.
View streaming output and exact model IDs
02Inspect usage
Review requests, token usage, and cost by API key.
Includes request status and latency
03Compare popular models
See each model's share of requests over the last 7 days.
Compare share and movement over time
Playground · Usage · Rankings
Use your client's native API format
Choose the protocol your client or SDK sends, then set the matching Base URL, API key, and model ID.
Connection settings
Recommended for: OpenAI SDKs, Codex, agents, and new applications
SDK base URL
https://router-api.xiu.ai/v1
Request path
/responses
Authorization header
Authorization: Bearer $XIUROUTER_API_KEY
cURL · request example
curl https://router-api.xiu.ai/v1/responses \
-H "Authorization: Bearer $XIUROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"gpt-5.5","input":"Hello"}'Full documentation: docs.xiu.ai/router/quickstart
Make your first request
Add funds, convert them to router credit, create an API key, and send a request.
- 01
Add funds
Choose an amount and payment method.
- 02
Convert to router credit
Move wallet balance to the credit used for API requests.
- 03
Create an API key
Set the models this key can call.
- 04
Send a request
Use the matching Base URL, API key, and model ID.
- 05
Check usage
Review the model, tokens, and cost.
Choose the right next step
Start with your use case, API format, or question, then continue to pricing, docs, or the console.
Start with a use case
Find guidance for agent development, OpenAI-compatible migration, multi-model apps, or usage reconciliation.
Choose an API format
Compare Responses, Chat Completions, Anthropic Messages, and Gemini before you integrate.
Browse the FAQ
Find answers about API keys, models, pricing, usage, request failures, and production checks.
Frequently asked questions
View all questions- Which API protocols are supported?
- XiuRouter supports OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, and Gemini GenerateContent. The protocol depends on your tool and model; use the matching Base URL and API key, then verify the model ID.
- How do service tiers differ?
- Service tiers differ in available models and pricing. With one available tier, read its prices directly. With several, use the tier assigned to your API key. Check service announcements before continuing to use a legacy group.
- How are savings calculated?
- For the same model, we compare the provider's reference input price with the XiuRouter input price for each service tier. Both use USD per 1M tokens. Savings vary by model, tier, and price type.
- Which models are supported?
- XiuRouter provides APIs for Claude, GPT and other models. Check Models and pricing for currently available model IDs, then confirm access with your API key. Supporting a request format does not mean every model from that format's developer is available.
- How is actual cost calculated?
- Billing uses the model, service tier, and actual input, output, and cache usage. Fast ×2.5 means Fast is priced at 2.5× Standard; Usage shows the actual service tier.
- Does XiuRouter store conversation content?
- XiuRouter does not store prompts, message history, files, images, or model responses, use conversations for training, or sell any user data. Usage statistics and model performance data are retained long term for billing, historical usage queries, and performance analysis. Model developers process requests under the data policies of their respective services.
Start with XiuRouter
Create an API key and make your first request
Choose a model and service tier, then use the matching Base URL in your client.