Module 1 · Foundations · Lesson 1 of 3 · 8 min
Model landscape
The models worth knowing in 2026, and which to pick for which job.
Video not playing? Some creators disable embedding.Watch on YouTube ↗
Lesson
The frontier is a small club: OpenAI (GPT-5.x), Anthropic (Claude 4.x), Google (Gemini 3), and a handful of open-source models. Every 3–6 months rankings shift; use whichever leads on your task today.
Task-to-model routing: long-form writing → Claude for tone, GPT-5 for structure. Data analysis → Claude or GPT-5. Image → Gemini or GPT-Image. Video → Sora, Veo, or Runway.
Do not marry a model. Use an abstraction layer (OpenRouter, Portkey, or a router library) so you can swap without rewriting your workflows.
Key takeaways
- ✓3 frontier providers + open source cover 95% of needs.
- ✓Task-to-model routing beats loyalty to one provider.
- ✓Use an abstraction layer to future-proof.
Exercises
- 1Try the same prompt on Claude, GPT, and Gemini. Note which handles it best. That's your task-to-model mapping.