Fetched May 1, 2026
Introduced the new Flex and Priority inference tiers, offering more options for optimizing cost or latency.