AI Playbook
AI Playbook
How to combine apps and models, how to get tokens and quotas, and how to spend less. Prices and limits in these guides were checked against official pages, with the check date on each article.
Draft on a cheaper model, review on a stronger one
Send high-volume, reversible work to a cheaper or free-tier model, and reserve a stronger model for hard, outward-facing, or irreversible steps. Prices below were checked against official pages on 10 Nov 2026.
How to read official API prices: start with per-million-token rates, then cache and tiers
Use each provider’s official pricing page. Note input, output, and cache rates, and keep standard, batch, and free tiers distinct. Figures below were copied from official pages on 10 Nov 2026.
How to get tokens and quotas: free tiers, prepaid credit, and subscriptions
A chat-app subscription and a developer API balance are separate accounts. Free tiers, prepaid credit, and monthly plans are all opened in official product consoles; limits are whatever the page says that day.
Subscription or pay-as-you-go API: price the way you actually work
A subscription suits frequent use inside the official app when you do not want to count tokens. The API suits your own tools, or highly variable volume. The comparison below uses official list prices only and does not promise which will be cheaper for you.
Spend fewer tokens when you combine tools: cache, keep context short, and batch the work that can wait
Put repeated system instructions and knowledge bases on official prompt caching, attach only the material the step needs, and send jobs that can wait through Batch. The price gaps below come from official pages checked on 10 Nov 2026.
After Fable 5.2’s price cut: keep Opus for daily work, use Fable for harder jobs, and note the extra usage window
Anthropic’s official guidance is to start most coding and knowledge work on Claude Opus 5.5, move up to Fable for harder jobs, and move down to Sonnet 5.5 or Haiku 5.5 for simpler delegated work. Fable 5.2’s standard API price is lower than Fable 5.1. Time-limited Fable usage on paid plans follows the Help Center; figures were checked on 10 Nov 2026.
Before Gemini’s promo rates rise: Argon goes up at the end of November, Flash at year end
Google’s AI Studio pricing page lists Gemini 4 Argon as paid-only, at $2 input / $10 output per million tokens through 30 Nov 2026, then $4 / $20 from 1 Dec 2026. Gemini 3.8 Flash’s paid promotional rates last through 31 Dec 2026 and rise on 1 Jan 2027. From 1 Dec 2026, requests to Gemini 2.5 Pro and 2.5 Flash are rerouted to newer models and billed at those models’ rates. Figures were checked on 10 Nov 2026.