Pricing you can read
You pay for the tokens you use, and every model lists its own price.
Early access
Bring your application. We handle the inference.
SnowQuery AI gives you affordable access to open-weight models through an API in the OpenAI format. The models run on our own hardware, so there are no GPUs to buy and no model servers to run.
chat.pyExample
# Keep your OpenAI client. Change the base URL.
from openai import OpenAI
client = OpenAI(
base_url="https://api.snowquery.ai/v1",
api_key=SNOWQUERY_API_KEY,
)
reply = client.chat.completions.create(
model="your-chosen-model",
messages=[{"role": "user", "content": "Hello"}],
)An API you already know
The API follows the OpenAI format, so existing clients, SDKs, and tools work after a base URL change.
Clear model information
Each model states what it is good at, its context length, and its limits.
Models
A focused catalog that grows with demand
We are starting with a small set of open-weight models and limited capacity, and adding more as usage grows. The launch catalog is being finalized. Each model will be listed here with:
- What the model is good at
- Context length and limits
- Price per million tokens
Pricing
Pay for what you use
Billing is by usage, counted in tokens. Prices are set per model and published here when access opens.
Custom models
Have a custom or fine-tuned model?
If you have a model of your own that you want hosted, we are glad to talk it through. Every model we serve runs on hardware operated by SnowQuery.
Request access
Access opens in stages
Capacity is limited while we grow, so we are opening access a few teams at a time. Requests are not open yet. Check back here soon.