Claude Opus 5.5 cuts cached-token read prices by 60%
Anthropic says Opus 5.5 input and output token prices fell 20% for token-billed usage, while cached-token reads fell 60%. Claude Code users can also set effort levels without resetting cache, and API-key and cloud-provider users can set a one-hour cache lifetime.
SessionWatcher editorial · Published · Updated · Source announcement: 2026-09-24
Related guide: How to check your Claude Code usage
Opus 5.5 cached reads cost less
Anthropic says that for usage billed by the token, Opus 5.5 input and output token prices were reduced 20%, and the price of reading a cached token fell 60%. The post says cache reads make up the majority of agentic and coding work costs.
The post frames the price change around longer, context-heavy coding sessions. Anthropic says input that misses the cache decreased by more than 50%, and that cache reads account for a larger share of today's Claude Code traffic than six months earlier.
Sources: Coding sessions are longer and use more context. Claude Opus 5.5 is built with that in mind.
Cache behavior in Claude Code
Anthropic says users of newer models such as Opus 5.5 can change effort levels during a session without resetting cache. Developers using API keys and cloud providers can set a one-hour cache lifetime, which Anthropic says subscribers already had. Forked subagents also start from the parent's cache rather than paying for the same context again.
The post recommends running /usage in Claude Code to see how much usage is cached reads. It also advises choosing a model at the start of a session rather than switching midway, and compacting before stepping away.
Sources: Coding sessions are longer and use more context. Claude Opus 5.5 is built with that in mind.
AI assisted reporting, checked against the linked official sources. Source pages checked 2026-10-03. Editorial process and corrections.