The point: Model selection determines the available intelligence, while effort level controls how intensively Claude works—such as how many files are read, tests executed, and steps processed.
Anthropic distinguishes between model capacity and effort level in Claude Code—two control mechanisms that serve different functions and must be deployed strategically to achieve optimal results.
Claude Code offers two independent settings: model selection (such as Claude Fable 5 or Claude Sonnet) and effort level. These control different aspects of how Claude works. The model defines the weights and thus the fundamental performance capability; larger models are demonstrably more capable than smaller ones according to industry benchmarks.
Effort level, however, is more than mere “thinking time.” It controls how much total effort Claude expends on a request: how many files are read, how thoroughly checks are performed, and how far Claude progresses through a multi-step task before requesting feedback. At higher effort levels, Claude reads more files, runs tests, and verifies results before providing the response. At lower effort levels, Claude would sooner ask for additional context rather than consume tokens itself.
For practical application: smaller models are sufficient for routine tasks, larger models are worthwhile for complex or ambiguous problems. The effort level should be set as a general default based on the type of regular work, not adjusted for each individual task. If Claude fails despite having complete context, that signals the need for a more capable model. Conversely, if Claude misses a file, fails to run tests, or aborts a refactoring process, a higher effort level helps.
Source: claude.com · Published July 6, 2026
Lumi AI News — AI-assisted curation pursuant to Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.3.