Struct BenchmarkTask 

pub struct BenchmarkTask {
    pub name: &'static str,
    pub prompt: &'static str,
    pub max_steps: usize,
    pub verify: fn(&BenchmarkResult) -> f64,
}

A single benchmark task.

Fields

name: &'static str

Short name

prompt: &'static str

Prompt sent to agent

max_steps: usize

Max steps budget (Karpathy: fixed budget per experiment)

verify: fn(&BenchmarkResult) -> f64

How to verify success — function checks the output

Trait Implementations

impl Clone for BenchmarkTask

fn clone(&self) -> BenchmarkTask

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for BenchmarkTask

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

Formats the value using the given formatter.