September 27, 2026
Claude Code Limits Go Further: Opus 5.5 Raises Them by 25%
On Sep 25, Anthropic said that Pro, Max, and Team limits with Opus 5.5 go about 25% further than with Opus 5. In the API, input fell from $5 to $4 per million tokens, output from $25 to $20, and cache reads from $0.50 to $0.20.
Addy Osmani
@addyosmani
Glad Claude limits feel different now! The lower price of Opus 5.5 directly increases your limits. You can get about 25% more done with them than with Opus 5. Cache reads cost 60% less, and the model produces output 30% faster: https://claude.dev/blog/what-a-task-costs-on-opus-5-5/
Not sure what happened to Claude limits they went from barely usable to basically unlimited 😂
· 3.7K views
Previously, a long task in Claude Code would consume a shared subscription limit faster. With Opus 5.5, the same limit should go about a quarter further, while the model generates output more than 30% faster.
Switching in the CLI. Update Claude Code with `claude update` to version 2.1.280 or later, then select Opus 5.5 via `/model`.
Cache is cheaper. For tasks that retain a large context, cache reads now cost 60% less: $0.20 instead of $0.50 per million tokens.
Fast mode makes the API up to 2.5 times faster, but costs $8 per million input tokens and $40 per million output tokens.
