Models and suppliers — Omnia Model Hub
The live marketplace is served from the platform database, so the website and the console never disagree.
What a listing contains
Each row is a supplier's offer for one model in one region. It records input, output and cached-input prices per million tokens, first-token latency, throughput, seven-day success rate, 429 rate, tool-call quality, data policy, zero-retention support, quantisation, service tier and trust grade.
Why the same model has several prices
Suppliers publish their own quotes. Buyers compare price and latency rather than being shown a single upstream list price. A request can restrict suppliers, require a data policy, cap price, prefer low latency or sort by price.
Live catalogue
The table below is rendered from the same endpoint used by the console. Use the console to create an API key and call a model.