PRICING

Intelligent Compute Platform Pricing

Model prices, capability tags, cache and tiered rules. Filter by vendor, tag, billing mode, and keyword.

Models

—

Vendors

—

With cache

—

Tiered

—

Loading…

Prices may change with model versions. Actual usage settlement governs.

Per-Token

Chat and inference models are billed by input/output tokens, each at its own unit price.

Cache

Some models support cache read/write. Cache read is typically far cheaper than fresh input; cache creation is billed at a one-off rate.

Tiered

Some models price by input context length tier. Open the card for input/output/cache unit prices per tier.

Model ID copied