A few days ago, Claude Fable 5 appeared in the model dropdown. Without warning.

What happened

We work daily with Claude Code – for social ops workflows, automation, content. We tested Fable 5 immediately. Same prompts, old sessions, familiar tasks.

The difference was significant. Better context, fewer repetitions, more complex relationships without much rework. For multi-step automation projects – where one model plans and others execute – a true leap forward.

Why it was brief nonetheless

Usage limits were quickly reached. Fable is more expensive to run than Sonnet or Haiku – which makes sense given the increased performance. But: we would have liked to experiment more.

We will see when Fable 5 becomes regularly available.