01
One unified endpoint
Use one OpenAI-compatible API instead of maintaining separate SDKs, credentials, and request formats for each provider.
One endpoint forevery model.
One API for teams building on AI. Models are served through vendors' own APIs and cloud model services, without rebuilding your integration for every provider.
Input / 1M
$1.7143
Official price $2.50
Output / 1M
$5.1429
Official price $7.50
178 models · 10 vendors · 2 channels
Official price checked 2026-10-23. Struck-through prices are Qwen's own official list price for this model, checked 2026-10-23. Qwen pricing page
1# Official OpenAI SDK - only base_url changes2from openai import OpenAI34client = OpenAI(5 base_url="https://www.realrelay.ai/v1",6 api_key="sk-realrelay-***",7)89response = client.chat.completions.create(10 model="deepseek-v4-flash",11 messages=[{"role": "user", "content": "Summarise this report"}],12)1314print(response.choices[0].message.content)Latest models
Recently released models available in our catalog.
- gpt-6.1-sol
openai/gpt-6.1-sol - claude-sonnet-5-5
anthropic/claude-sonnet-5-5 - claude-opus-5-5
anthropic/claude-opus-5-5 - gpt-6-sol
openai/gpt-6-sol - gpt-6-luna
openai/gpt-6-luna - gpt-6-astra
openai/gpt-6-astra - claude-fable-5-1
anthropic/claude-fable-5-1 - glm-5.3
z-ai/glm-5.3 - deepseek-v4.1-flash
deepseek/deepseek-v4.1-flash - glm-5.3-flash
z-ai/glm-5.3-flash - gpt-image-2.5-sunburst
openai/gpt-image-2.5-sunburst - gpt-image-2.5-flare
openai/gpt-image-2.5-flare
Why RealRelay.ai
Stop rebuilding AI infrastructure from scratch
Replace separate provider integrations with one API key and one familiar request format, with clear model choices, published prices, and routing options.
02
Published prices, one credit balance
Compare published model prices in one catalog, then use one prepaid credit balance across available models.
03
Change the model ID, not your application
Keep the familiar request format and move between supported models without rewriting your integration.
04
Routing you can control
When a model has multiple channels, choose the default, lowest-price, or lowest-latency strategy at the account level.
Credits
Pay for what you use
Prepaid credit, deducted per request at each model's published rate. No subscription and no monthly commitment.
Try it
01A first top-up to test the API, compare a few models and wire up an integration.
Top-up amount
$20
Pay as you go
02Top up what you expect to use and add more when you need it. Your balance is deducted only when a request is billed.
Top-up amount
From US$10
Teams
03API keys for each member or project, each with its own quota and model allowlist. SSO, an SLA and monthly invoicing are available under a written enterprise agreement.
Top-up amount
Let's talk
Credit-based, not a monthly subscription. Your balance is deducted only when you use it.
01
Getting started
What the gateway is and how an existing integration moves over.
Is RealRelay.ai a model provider?
No. RealRelay.ai is not a model provider. Instead, it connects to upstream model providers to offer reliable, stable, secure, and unified smart routing and model access. It supports multiple API formats compatible with mainstream providers, including OpenAI, Anthropic, and Gemini.
Do I have to rewrite an existing OpenAI integration?
Usually not. Point base_url at https://www.realrelay.ai/v1, swap in a RealRelay.ai key, and pick a model id from the catalog. The request shape stays the same.
Where does the catalog data come from?
From a catalog snapshot the operator publishes from the platform — the model page shows the date it was published. Struck-through original prices come from our channel list-price records; what we charge always comes from the platform.
02
Routing and billing
How channel choice, model capabilities, and usage charges work.
Can I pin a specific channel?
Yes. Write the bare model name to let the gateway pick the cheapest channel serving it, or prefix it with a channel name to lock a specific one. The routing strategy (default, lowest price, lowest latency) is an account-level setting in the console.
Are streaming and tool calling supported?
Streaming works via the standard stream parameter over SSE. Tool calling depends on the model. The catalog lists which models report it.
How is billing structured?
You prepay credits in the console wallet, and usage is deducted per million tokens at each model's published rate. Image and media models bill per call. The catalog shows the mode alongside the price.