The latest flagship AI assistant delivers performance parity with Fable 5.1 for most general tasks and operates 30% faster than the previous Opus 5 iteration. Anthropic priced the model at $4 per million input tokens and $20 per million output tokens, a 20% reduction from Opus 5, while cache reads dropped 60% to $0.20 per million tokens. The system is now the default for Pro, Max, and Team subscription plans, which include a 20% increase in 5 hour usage limits.
The launch introduces a safety mechanism where the model falls back to less capable versions for specific frontier LLM development tasks, including kernel coding. In enterprise trials by Box, the model used 63% fewer tokens and reduced verbosity by 42% for knowledge work tasks. General availability has extended to third party integrations such as GitHub Copilot, Cursor, and Devin.
Key sources
- SOURCE@claudeai“costs 40% less to run than Opus 5”x.com
- SUPPORT@_catwu“Your rate limits will go 25% further on Opus 5.5 compared to Opus 5”x.com
- SUPPORT@bcherny“Opus 5.5 finished in 9.5 hours compared to Fable 5.1's 12 hours, and for 51% less cost”x.com
- SUPPORT@m1astra“first Opus meant to fall back to a less capable model for "a small set of capabilities related to the development of frontier LLMs, such as kernel development.."”x.com
- SUPPORT@stocksavvyshay“delivers Claude Fable 5.1-level performance on most tasks while costing 40% less to run than Opus 5”x.com
- SUPPORT@levie“63% fewer tokens used, 42% less verbosity, and 30% faster vs. Opus 5”x.com
- SUPPORT@zephyr_z9“puts the most important information up front and follows the writing rules you give it”x.com
- SOURCEmarketbrief.now