token-efficiency

To reduce token usage, apply token-efficient mode to minimize unnecessary tool calls, combine queries, reuse context, and select the cheapest tools while preserving result accuracy.

The best way to minimize tool calls during multi-step data retrieval is to apply context-aware optimization, which combines queries and reuses information to reduce overhead and enforce targeted access for reads.

Token-efficient optimization does not compromise result quality; it enforces guidelines to choose the most appropriate tools, minimize reads with targeted access, and deliver concise results while operating in a cost-effective mode.

You should use token efficiency guidelines whenever token costs matter, including during long-running sessions, searches, data retrieval, and multi-step workflows where minimizing overhead and unnecessary tool calls is essential.

Multi-step AI workflows consume excessive tokens during long-running sessions due to unnecessary tool calls and unoptimized reads; applying token-efficient mode enforces targeted access and context reuse to minimize this overhead.