cli.test_cmd()

Run eval suite for a persona across one or more models.

Usage

Source

cli.test_cmd(
    persona,
    model,
    judge,
    output,
    threshold,
    no_guards,
)

Evaluates the persona’s test queries against each model, scores with a judge, and prints a scorecard.

Examples

talk-box test code_reviewer -m anthropic:claude-sonnet-4-6

talk-box test customer_support_tier1 -m ollama:llama4 -m anthropic:claude-sonnet-4-6 -t 0.85