Qtum AI Router
One key. Every model.
The Qtum AI Router puts text, image, audio and video models behind a single OpenAI- and Anthropic-compatible endpoint. Point your existing SDK at it, keep your code, and stop maintaining an account and an invoice per provider.
Why route through one endpoint
Four things change the day you stop integrating providers one at a time.
Your SDK already speaks it
The router implements the OpenAI Chat Completions API and the Anthropic Messages API. Change the base URL and the key — not your code. Streaming works over standard server-sent events.
One key, one balance
One account, one API key and one balance across every model and modality. No per-provider signup, no separate invoice, no juggling five dashboards to find out what a feature cost you.
Every modality on one endpoint
Text and vision, image generation and editing, text-to-speech and transcription, async video generation, and non-generative decision models — all addressed by model id on the same base URL.
You are charged for what ran
Each request reserves worst-case credits, settles against the usage the model actually reported, and refunds the difference. Every charge is a line item you can read back, down to the request.
Quick start
Two changes to an existing OpenAI integration: the base URL becomes https://router.qtum.ai/v1, and the key becomes your Qtum API key.
curl https://router.qtum.ai/v1/chat/completions \
-H "Authorization: Bearer $QTUM_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-5.5",
"messages": [
{ "role": "user", "content": "Explain Qtum AAL in one paragraph." }
]
}'curl https://router.qtum.ai/v1/messages \
-H "Authorization: Bearer $QTUM_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "claude-opus-4-7",
"max_tokens": 1024,
"messages": [
{ "role": "user", "content": "Explain Qtum AAL in one paragraph." }
]
}'Every modality, one base URL
Models are addressed by id on the endpoint that matches their modality.
Text & vision
/v1/chat/completions · /v1/messagesChat and vision models priced per million input and output tokens, with streaming and both the OpenAI and Anthropic request shapes.
Image
/v1/images/generations · /v1/images/editsGeneration and editing. Billed per generated image, or per token for the token-metered models.
Audio
/v1/audio/speech · /v1/audio/transcriptionsText-to-speech and transcription, billed per token, per million characters or per audio second depending on the model.
Video
/v1/videosAsynchronous generation: create a task, poll it, then fetch the result. Billed per output second or per completion token.
Decisions
/v1/decisionsNon-generative classification and scoring models that return a decision rather than prose, billed per request, per decision or per input token.
Pricing
Usage is billed in credits. 1000 credits equal 1 USD of deposit, so a credits figure divided by 1000 is the dollar figure.
Text models are priced per million input and output tokens. Image models are priced per generated image, or per token where the model meters tokens. Audio is priced per token, per million characters or per audio second. Video is priced per output second or per completion token. Every request reserves the worst case up front, settles against the usage the model reported, and refunds the difference — so a short answer from an expensive model costs what the short answer costs.
A sample of the text catalog
| Model | Provider | Input / 1M tokens | Output / 1M tokens | Context |
|---|---|---|---|---|
| kimi-k2.6 | Moonshot AI | 1 cr$0.0012 | 1 cr$0.0012 | 128K |
| gpt-6-luna-ab | OpenAI | 10 cr$0.0096 | 48 cr$0.05 | 272K |
| qwen3.8-27b | Alibaba | 10 cr$0.01 | 80 cr$0.08 | 262K |
| gpt-5.6-luna-ab | OpenAI | 19 cr$0.02 | 115 cr$0.12 | 105K |
| deepseek-flash-volc | DeepSeek | 71 cr$0.07 | 282 cr$0.28 | 1.02M |
| gpt-6-luna | OpenAI | 84 cr$0.08 | 420 cr$0.42 | 1.05M |
| gemini-3.8-flash-ab | 90 cr$0.09 | 450 cr$0.45 | 1.05M | |
| gpt-5.4-mini-ab | OpenAI | 90 cr$0.09 | 540 cr$0.54 | 400K |
Questions
What is the Qtum AI Router?
The Qtum AI Router is a single API endpoint that serves many AI models — text, image, audio and video — through one account, one API key and one balance. It is OpenAI-compatible and Anthropic-compatible, so existing SDKs and tools work against it without code changes.
Is it compatible with the OpenAI API?
Yes. Point any OpenAI client at https://router.qtum.ai/v1 with your Qtum API key as the bearer token and call /chat/completions as usual, including streaming. The images and audio endpoints follow the same OpenAI shapes.
Can I use it with the Anthropic SDK or Claude Code?
Yes. The router also implements the Anthropic Messages API at https://router.qtum.ai/v1/messages. For Claude Code and compatible tools, set ANTHROPIC_BASE_URL to https://router.qtum.ai and ANTHROPIC_AUTH_TOKEN to your Qtum API key.
Which models are available?
The catalog spans text and vision models, image generation and editing models, text-to-speech and speech-to-text models, video generation models, and decision models. It changes as upstream availability changes, so the live list and current prices are published on the Models page.
How is usage billed?
Usage is billed in credits, where 1000 credits equal 1 USD of deposit. Text models are priced per million input and output tokens; image models per image or per token; audio per token, per million characters or per audio second; video per output second or per completion token. A request reserves the worst case, settles on reported usage, and refunds the rest.
How do I get an API key?
Sign in to the Qtum AI console with a wallet, a Google account or an email code, open the API Keys page and create a key. The key authenticates to the router; it is separate from the credentials the router uses upstream.
How do I add credits?
Deposit QTUM or a supported stablecoin to the address shown in the console. Stablecoin deposits credit at 1000 credits per USD; QTUM deposits convert at the rate recorded at the time of the deposit.
Does it support streaming responses?
Yes. Streaming uses standard server-sent events, identical to the OpenAI and Anthropic streaming formats, so existing stream handling works unchanged. Token usage is reported in the stream and is what the request settles against.
Guides
Start with one key
Sign in, create an API key, and point your existing client at https://router.qtum.ai/v1.