Enum ThinkingAccounting 

pub enum ThinkingAccounting {
    InsideOutputBudget,
    SeparateFromOutput,
}

How a provider’s wire accounts reasoning/thinking tokens against the request’s output budget (I1, CP-12 review). The reserve model must count thinking exactly once, and provider differences are represented explicitly instead of being summed blindly.

Variants

InsideOutputBudget

Thinking spends from the request’s output cap: Anthropic messages (thinking.budget_tokens < max_tokens), OpenAI Responses/Chat and Codex wires (reasoning tokens count toward max_output_tokens / max_completion_tokens). Reserve ONE envelope: max(output_reserve, thinking_budget).

SeparateFromOutput

Thinking is budgeted separately from response output (Gemini thinkingBudget), or the wire is unknown — reserve both envelopes, which errs toward earlier compaction, never toward exhaustion.

Trait Implementations

impl Clone for ThinkingAccounting

fn clone(&self) -> ThinkingAccounting

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for ThinkingAccounting

impl Debug for ThinkingAccounting

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

Formats the value using the given formatter.

impl Eq for ThinkingAccounting

impl PartialEq for ThinkingAccounting

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for ThinkingAccounting