This is an actual f*cking goldmine
Anthropic just shared one of the best tricks for Claude Code. The idea: run Opus 5.5 and three Sonnet 5.5 subagents, each one keeping its own memory for the job it's best at. Less repeated explaining, because every task lands on an agent that
How to run sonnet as lead haiku as swarm and opus as advisor on call
Anthropic just shipped the cleanest multi-model split in Claude Code. one command, three models, way fewer tokens. the play: sonnet 5.5 drives the session > haiku 5.5 swarms search/grep/audit in parallel >
Claude Code ran 18% leaner than Codex on our ticket.
It also left the fix incomplete 7 of 12 times. Claude Code 2.1.317 just shipped richer cost reporting. So we ran one real support ticket 12 times on each agent, same prompt, graded against 7 hidden tests. Claude Code (Opus
Holy shit! Brothers and sisters, if you want to make videos with Opus 5.5, you absolutely have to bookmark these 8 GitHub repos.
1. awesome-opus5-5-videos (about 7100 stars) The largest collection of hit videos right now, with the prompt behind each one included.
You have reached the end of the archive
All of Claude Opus 5