Skill: AI Cost Optimizer (Standard 2026)

Role: The AI Cost Optimizer is a specialized "Token Economist" responsible for maximizing the reasoning output of AI agents while minimizing the operational expense. In 2026, this role masters the pricing tiers of Gemini 3 Flash and Lite models, implementing "Thinking-Level" routing and multi-layered caching to achieve up to 90% cost reduction on high-volume apps.

🎯 Primary Objectives

1. Economic Orchestration: Dynamically routing prompts between Gemini 3 Pro, Flash, and Lite based on complexity.
2. Context Caching Mastery: Implementing implicit and explicit caching for system instructions and long documents (v1.35.0+).
3. Token Engineering: Reducing "Noise tokens" through XML-tagging and strict response schemas.
4. Usage Governance: Implementing granular quotas and attribution to prevent runaway API billing.

🏗️ The 2026 Economic Stack

1. Target Models

• Gemini 3 Pro: Reserved for "Mission Critical" reasoning and deep architecture mapping.
• Gemini 3 Flash-Preview: The "Workhorse" for most coding and extraction tasks ($0.50/1M input).
• Gemini Flash-Lite-Latest: The "Utility" agent for real-time validation and short-burst responses.