A: Hey mrt,

Good question! We don't have a single fixed rate card for all models via the Pixel ML OpenRouter because it works primarily as a pass-through billing system.

Here’s how the cost breaks down when you use it:

Base LLM Cost: You pay the standard rate that the underlying LLM provider (like OpenAI, Anthropic, Google, etc.) charges for the specific model you use (e.g., GPT-4o, Claude 3 Sonnet). This is the core cost.
Small Overhead: We add a minimal fee on top of the base LLM cost to cover:
Taxes: As a registered corporation (C-corp), we need to handle applicable sales taxes.
Transaction Fees: Payment processing fees from services like Stripe.
Why use OpenRouter then?

Convenience: You don't need to manage multiple API keys and billing relationships with different LLM providers.
Access & Rate Limits: It can give you access to models or higher rate limits that might require significant pre-spending directly with the provider (e.g., some heavy tasks like DeepResearch might need an OpenAI key with Tier 4 spending history, which our OpenRouter can facilitate).
In short: The cost is essentially the LLM provider's public price + a small percentage for tax/transaction fees. It's designed for convenience and access, not as a separate profit markup on the AI calls themselves. To estimate, check the public pricing page of the specific LLM provider for the model you plan to use.