Struct ReasoningBudget 

pub struct ReasoningBudget {
    pub samples: usize,
    pub refine_rounds: usize,
}

Compute budget for a reasoning-max run. Higher = more model calls = higher reliability at higher cost/latency.

Fields

samples: usize

Number of independent drafts to sample before selecting (best-of-N).

refine_rounds: usize

Number of critique→revise rounds applied to the selected draft.

Implementations

impl ReasoningBudget

pub fn for_tier(tier: &TaskTier) -> Self

Default budget for a task tier. Cheap tiers get a single greedy pass; hard tiers get search + iterative refinement.

pub fn scaled_for_capability(self, band: u8) -> Self

Weaker models benefit MORE from test-time compute — they have more headroom to recover from a bad first sample. Spend more on them. band is the model capability band (1 = small … 5 = frontier); see model-capability-profile.json.

pub fn max_calls(&self) -> usize

Rough upper bound on model calls this budget will issue, for cost gating.

Trait Implementations

impl Clone for ReasoningBudget

fn clone(&self) -> ReasoningBudget

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for ReasoningBudget

impl Debug for ReasoningBudget

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for ReasoningBudget

impl PartialEq for ReasoningBudget

fn eq(&self, other: &ReasoningBudget) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for ReasoningBudget