Every open model.
One API. Pay in crypto.
DeepSeek, Kimi, Qwen, GLM and Mistral behind one OpenAI-compatible endpoint. Top up with USDT, USDC or BTC — no card needed, credit lands in minutes.
from openai import OpenAI client = OpenAI( base_url="https://api.zardon.xyz/v1", api_key="zd-...", ) resp = client.chat.completions.create( model="deepseek/deepseek-v4-pro", messages=[{"role": "user", "content": "Hello!"}], ) # swap the model string for kimi, qwen, glm…
Three steps, no bank involved
Your code doesn't change. Point any OpenAI-compatible SDK, agent framework or IDE plugin at Zardon.
Top up with crypto
Send USDT, USDC, BTC or ETH to your deposit address. Balance is credited after network confirmation.
Create an API key
One key works for every model. Set per-key spend limits so an agent loop can't drain your balance.
Call any model
Pay per token from your prepaid balance. No subscription, no minimum, unused credit never expires.
Models & prices
Per 1 million tokens, in USD, deducted from your balance. Hundreds more models available — these are the popular ones.
| Model | Context | Input | Output |
|---|---|---|---|
| DeepSeek V4 Prodeepseek/deepseek-v4-pro | 1M | $0.99 | $1.65 |
| DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flash | 1M | $0.055 | $0.22 |
| Kimi K3.1 Promoonshotai/kimi-k3.1-pro | 1M | $0.26 | $9.24 |
| Qwen4 Maxqwen/qwen4-max | 1M | $1.57 | $4.70 |
| Qwen3.8 Flashqwen/qwen3.8-flash | 1M | $0.165 | $0.52 |
| GLM-5.3z-ai/glm-5.3 | 1M | $0.077 | $1.54 |
| GLM-5.3 Flashz-ai/glm-5.3-flash | 1M | $0.044 | $0.086 |
| Mistral Large 4mistralai/mistral-large-4-0 | 1M | $0.77 | $2.31 |
Prices may change when model providers change theirs; the price at the time of each request applies.
Built for people who'd rather not use a card
Card declined abroad, a provider that won't take your bank, or you just keep your budget in stablecoins — Zardon is for you.
OpenAI-compatible
Chat completions, streaming, tool calling and JSON mode. Works with LangChain, LlamaIndex, Cline, Continue, Aider and anything that speaks the OpenAI API.
Automatic fallback
If one provider is slow or down, requests route to another host of the same model. Your agent keeps running.
Prepaid, so no surprises
You can never be billed more than your balance. Per-key limits and usage logs per model.
Agent sandboxes & GPU jobs Soon
Run code your agents write in isolated sandboxes, and rent GPUs by the minute — paid from the same balance. Tick the box below to get early access.
FAQ
Which coins and networks do you accept?
USDT on TRC-20 (Tron) and ERC-20 (Ethereum), USDC on Ethereum and Base, BTC and ETH. Minimum top-up is $5.
How fast is my balance credited?
As soon as the network confirms: usually 1–3 minutes for USDT on Tron, about 10–30 minutes for BTC.
Do I need a credit card or bank account?
No. Sign up with an email, top up with crypto, and you're ready.
Is my data used for training?
We don't store prompt or completion content beyond what's needed to bill you. Each model provider's own data policy applies to requests routed to it.
Can I get a refund of unused balance?
Unused credit never expires. Refunds of unused balance are available within 30 days of a top-up, minus network fees.
Get early access
We're onboarding accounts in batches. Leave your email and we'll send your API key and $5 of free credit.