How to Prompt Claude Fable 5.2 [Anthropic's Own Guide]
Checkout ChatLLM https://chatllm.abacus.ai/rsh Claude Fable 5.2 is out, and Anthropic still says to keep Opus 5.5 as your daily driver. In this video I go through the model highlights, the price and the plan limits, and Anthropic's new playbook for picking a model: when to use Fable, Opus, Sonnet or Haiku, which effort level to use in Claude Code, and why to change the effort before you change the model. Then I go through Anthropic's prompting guide for Fable 5.2 as five changes: describing outcomes instead of steps, how to stop it ending a job early, the one sentence about code ownership, keeping subagents in check, and how exact your rules and approvals now need to be, including what to remove from prompts you wrote for Opus. COURSE & COMMUNITY https://switchdim.s.gy/yt-fable-vs-opus SOCIALS X: https://x.com/robshocks LinkedIn: https://www.linkedin.com/in/robertoshaughnessy/ Instagram: https://www.instagram.com/therobshocks/ WORK & COLLABS info@switchdimension.com CHAPTERS 0:00 Intro 0:43 Fable 5.2 Highlights 1:43 Which Model for Which Job 3:02 Change the Effort Before the Model 5:18 Sponsor: ChatLLM 6:11 Prompting 1: Outcomes, Not Steps 7:22 Prompting 2: Make It Finish 8:30 Prompting 3: Say Who Owns the Code 9:24 Prompting 4: Keep Subagents in Check 10:50 Prompting 5: Be Exact About Rules 11:41 When I'd Use Each 13:18 Wrap Up #ClaudeFable #Fable52 #ClaudeOpus #Opus55 #ClaudeCode #Anthropic #AIDevelopment #AgentDevelopment #PromptEngineering #AICoding #ClaudeAI #AIModels #CodingAgents #SpecDrivenDevelopment
