Offload a self-contained subtask to a cheaper model and return its output.

WHEN TO USE: hand off replaceable grunt work to save your premium tokens — boilerplate code, small bug fixes, formatting, translation, reading/summarizing long documents, drafting routine copy. Do NOT delegate judgment work (planning, architecture, final review, talking to the user) — keep that for yourself.

The delegated model sees ONLY the task string and has NO access to this conversation. So make task fully self-contained (include all needed context).

Args: task: Complete, self-contained instruction for the cheap model. model: "auto" (default) routes by task — code→deepseek, long docs→kimi, Chinese→qwen, multi-step→mimo, quick chores→flash. Or force an alias (deepseek/mimo/flash/kimi/qwen) or a raw proxy model_name. max_tokens: Output cap. Default 4000 (kept large so reasoning models that spend budget on hidden thinking still return non-empty text).

Note: a call typically takes ~10-60s (longer for big inputs) and blocks until the cheap model returns, so prefer one focused task per call.

Returns: The model's text output, or a string starting with "[delegate-error]" on failure.

Show what delegating has kept off your premium quota (estimates, list prices).

List the model aliases available to delegate(), with their best use.