M03 · Understanding LLMs & the Model Landscape

Stage B — Literacy · Persona: Everyone · Duration: 90 min · Format: live / remote · Prereqs: none (M01 helpful)

Objective

Demystify what an LLM is, map the model landscape (Claude vs. GPT vs. Gemini vs. open models), and give the team a practical way to choose the right model for a task.

Why it matters

People waste money and trust by using the wrong model — a heavyweight model for trivial work, or a small one for work that needs reasoning. Literacy here makes every later module land better.

Learning outcomes

By the end, participants can: - Explain in plain language what an LLM is and how it differs from search or traditional software. - Name the major model families and what each is good at. - Distinguish model tiers (e.g., Opus / Sonnet / Haiku) and pick one for a given task. - Understand the basics of context windows, tokens, and why cost/speed/quality trade off. - Apply the company model-choice cheat sheet to real tasks.

Agenda

1. What an LLM actually is (and isn't) (15)
2. The landscape: Claude, GPT, Gemini, open models — strengths & where each fits (20)
3. Tiers within a family: when to use Opus vs. Sonnet vs. Haiku (15)
4. Context windows, tokens, cost/speed/quality trade-offs — practically (15)
5. Lab: route 8 real tasks to the right model (15)
6. Build the company cheat sheet (10)

Deliverable

Company model-choice cheat sheet → deliverable.md. Adopted by the team; lives in the Readiness File.

Gate contribution

Gate 2 (Literate).