Evals

Budgeted evaluation suite with keyword, length, and LLM judge gates. Numbers from the latest committed baseline.

Eval results chart
Executions154
Passed gates145
Avg combined score0.923
Spend$4.99 / $5.00
Provideropenrouter · anthropic/claude-sonnet-4
Single-agent142/142
Swarm4/12
Read full eval methodology