For those who have raised Claude Code to Opus 5.5,
Audit CLAUDE.md now It's better to let it out. Anthropic official release of Opus 5.5 Lowered the default effort to medium. If you keep the settings from Opus 5, Discard 20% of the tokens and keep moving. If you paste it in your todo-app (CLAUDE.md line 26), 1
Claude Sonnet 5.5 is out 👀
Anthropic just dropped a guide covering how to build with it, migrate from Sonnet 5, tune effort, use Claude Code, and choose between Sonnet 5.5 and Opus 5.5.
Opus 5.5 calls Grok to run much better than native open source. However
Opus 5.5 calls Grok to run much better than native open source. However, a 5-minute MTV consumes my weekly grok quota of a premium+ and a premium X account, and I have not finished producing the footage.
Anthropic released Sonnet 5.5 on September 28. The price remains unchanged: input $2, output $10/million tokens.
Official: To do the same work, tokens are much less, and each task is up to 30% cheaper; Terminal-Bench is 70.6%, exceeding Opus 5.5’s 66.4%. Third-party Artificial Analysis: At the highest level, each question costs about 193,000 tokens and $7.6, which is about half more expensive than Sonnet 5.
You have reached the end of the archive
All of Claude Opus 5