Effort
A parameter in large language model reasoning systems that controls the depth of internal chain-of-thought computation before an answer is generated. Commonly offered at Low, Medium, and High levels, it directly affects output quality, token consumption, and response latency.
Core Facts
Timeline (last 90 days)
Claude Code's Effort (thinking budget) setting can cause roughly a 11.9x cost difference for the same model on the same task.
For teams that invoke Claude Code in large volumes every day, using Effort tiers is a necessary cost-control practice.
In Claude Code, the way to view the current Effort level is to enter the Effort command, press S to change it only for this session, or press Enter to save it as the default.
In Claude Code, allocating Effort by task type can optimize cost: the Low level is sufficient for search-type subagents, while only refactoring and complex reasoning need High or higher levels.
In Claude Code, the Max Effort level only applies to the current session and does not automatically carry over to the next session.
In Claude Code, when a user specifies an unsupported Effort level (such as X High) for an older model, the system silently downgrades to the highest level that model supports (usually High) and continues execution, without error or notice.
The three ways Effort is implemented are: the definition layer (hard-coded in the Agent definition file), the cap layer (set an Effort ceiling in settings, with automatic downgrade when exceeded), and the audit layer (Tasks marks the model and level of each subagent).
The Effort override chain has four layers: environment variables can override per-call parameters; settings still take effect after Resume; definition file fields can lock in a level; and writing Effort in an Agent definition file can fix that agent to a low budget.
Claude Code allows specifying the Effort parameter individually for each subagent, replacing the previous mechanism where the entire session had only one global level.
Claude Code's official cost table shows the relative cost index of Effort levels with High as the baseline: Low is 0.47X, High is 1X, Max is 5.59X, and the deepest level is about 11.9X.
13 more timeline events
All Facts (20)
Haiku 5.5 is the first Haiku model with an adjustable effort setting.
70%UnverifiedFor teams that invoke Claude Code in large volumes every day, using Effort tiers is a necessary cost-control practice.
50%UnverifiedIn Claude Code, the way to view the current Effort level is to enter the Effort command, press S to change it only for this session, or press Enter to save it as the default.
50%UnverifiedIn Claude Code, allocating Effort by task type can optimize cost: the Low level is sufficient for search-type subagents, while only refactoring and complex reasoning need High or higher levels.
50%UnverifiedIn Claude Code, the Max Effort level only applies to the current session and does not automatically carry over to the next session.
50%UnverifiedIn Claude Code, when a user specifies an unsupported Effort level (such as X High) for an older model, the system silently downgrades to the highest level that model supports (usually High) and continues execution, without error or notice.
50%UnverifiedThe three ways Effort is implemented are: the definition layer (hard-coded in the Agent definition file), the cap layer (set an Effort ceiling in settings, with automatic downgrade when exceeded), and the audit layer (Tasks marks the model and level of each subagent).
50%UnverifiedThe Effort override chain has four layers: environment variables can override per-call parameters; settings still take effect after Resume; definition file fields can lock in a level; and writing Effort in an Agent definition file can fix that agent to a low budget.
50%UnverifiedClaude Code allows specifying the Effort parameter individually for each subagent, replacing the previous mechanism where the entire session had only one global level.
50%UnverifiedClaude Code's official cost table shows the relative cost index of Effort levels with High as the baseline: Low is 0.47X, High is 1X, Max is 5.59X, and the deepest level is about 11.9X.
50%UnverifiedClaude Code's Effort (thinking budget) setting can cause roughly a 11.9x cost difference for the same model on the same task.
50%UnverifiedThe Effort adjustment mechanism dynamically controls the depth of the model's reasoning chain, allowing low-Effort settings to respond quickly and high-Effort settings to trigger deeper reasoning, thereby balancing speed and accuracy.
50%UnverifiedEffort setting is an adjustment knob introduced by Anthropic in the extended thinking feature, allowing users to specify how much compute the model invests in internal reasoning before generating a response.
50%UnverifiedLow effort is suitable for deterministic tasks such as information retrieval and format conversion, while high effort is suitable for complex analysis requiring multi-step reasoning and weighing multiple solutions.
50%UnverifiedEffort control is a parameter provided at the Claude Opus 5.5 API level that allows specifying how much computational resource the model invests in reasoning at call time, with the default value being Medium.
50%UnverifiedFor everyday Q&A, the Medium or High tier of a lightweight model is sufficient, and in most cases it is unnecessary to go up to Max.
50%UnverifiedLightweight model + Max is suitable for large-scale batch tasks that are budget-sensitive, require high throughput, and can tolerate slight quality fluctuations.
50%UnverifiedEffort is an amplifier, while model tier is the capability ceiling; adding thinking budget to a smaller model cannot bridge the capability gap inherent to the model itself.
50%UnverifiedThe combination of a lightweight model plus maximum Effort can often approach the perceived performance of a flagship model plus medium Effort.
50%UnverifiedRaising Effort for simple tasks yields almost no benefit, while simultaneously driving up both cost and latency.
50%