Transparent pricing
Models & Pricing
Simple USD pricing per one million tokens. No monthly minimum and no model-specific subscription.
Every model on the gateway
CompareRates are USD per 1M tokens and follow upstream provider pricing, so they can move. The models.json feed carries the same figures with a version hash.
All prices are shown in USD and exclude VAT, sales tax and any other applicable taxes. Tax is calculated and added at checkout based on your billing country. Business customers in the EU who supply a valid VAT number are invoiced under the reverse charge.
Hubrium
Included with every account
- OpenAI Chat Completions and Anthropic Messages
- Streaming responses and usage logs
- Scoped API keys and spend controls
- Access to every enabled model family
Pricing, without the small print
How is usage calculated?+
Input and output tokens are metered separately for each request. The exact usage appears in your account logs.
Is there a monthly commitment?+
No. Launch accounts use prepaid, usage-based credit with no recurring minimum.
What are cached input tokens?+
Some models discount repeated prompt content when their cache is hit. The reduced rate only applies when the provider reports a cache hit.
Can prices change?+
Yes. Upstream pricing and exchange rates can move. The models page and machine-readable model feed show current customer rates.
Can I get a refund?+
Yes. Prepaid credit you have not spent is refundable — email support and we return the unused balance to the original payment method. Credit already consumed by completed API requests is not refundable.
Are prices inclusive of tax?+
No. Prices exclude VAT and any other applicable tax. Tax is added at checkout based on your billing country, and EU business customers with a valid VAT number are invoiced under the reverse charge.