Anthropic model for Cline
Try Claude Fable 5 in Cline
Claude Fable 5 is available in Cline. Choose Cline (usage-billing) or Anthropic as the provider, select Claude Fable 5, and use it for long-running work where preserving intent across many steps is critical.
Best for
- Long-running agent tasks with many dependent implementation steps
- Demanding reasoning over a large repository and specification set
- Complex work where capability matters more than latency or cost
Limitations
- Fable is a higher-cost, slower tier than Anthropic's speed-oriented models.
- Its safety classifiers can decline requests, so production workflows should handle refusals and define a fallback.
Set up in Cline
Choose the surface that fits your workflow: use the Cline IDE extensions, run Cline CLI in your terminal or CI, or configure the model in scripts and products with the Cline SDK. Then complete the provider-specific steps below.
- Sign in to your Cline account in the Cline IDE extension or CLI.
- Choose Cline (usage-billing) provider, or choose Anthropic provider and add your own API key.
- Select Claude Fable 5 from the model list and divide long work into explicit checkpoints that can be reviewed.
Example tasks
Use these as starting prompts and adapt them to your repository.
- Task 1: Plan and execute the staged replacement of this permissions system, preserving compatibility across three deploys.
- Task 2: Investigate the full incident timeline, trace the defect across services, and implement the approved remediation with tests and runbooks.
Frequently asked questions
What is Claude Fable 5 best for in Cline?
It is best for long-running agents, demanding reasoning, and complex multi-step coding work where maximum capability is the priority.
Should every Cline task use Claude Fable 5?
No. Sonnet or Haiku can be more appropriate when speed and cost matter more than maximum capability.