Lesson 1005 of 1455
Reasoning Models (o1, o3, Claude Thinking) vs Regular Chat Models
Reasoning models 'think' before answering — slower and pricier, but way better on math, code, and logic.
Builders · Model Families · ~4 min read
The big idea
OpenAI's o1, o3, Claude with extended thinking, Gemini Thinking, DeepSeek R1, and Grok with reasoning — these all share a trick. Before answering, they generate a long internal 'thought' process. You don't see most of it. The result: way better on math, science, and tricky code. The cost: 10–100x more tokens and longer waits. Use them when accuracy matters.
Some examples
- Math word problems: reasoning models go from ~50% to ~90% correct.
- Hard coding tasks (fix this bug across 5 files): reasoning models clearly win.
- Asking 'what's the capital of France?': reasoning model is wasteful — use a regular model.
- Writing creative fiction: reasoning models are often worse — too cold and analytical.
Try it!
Run the same competition-math problem on a regular model and a reasoning model. Compare answer and time.
Key terms in this lesson
End-of-lesson quiz
Check what stuck
8 questions · Score saves to your progress.
Lesson help
Questions are best handled with a grown-up here.
For this age range, Tendril keeps freeform AI chat paused until parent/guardian consent and child-safe moderation are fully verified. Use the quiz, notes, and related lessons below, or ask a parent, guardian, teacher, or librarian to work through the question with you.
Progress saved locally in this browser. Sign in to sync across devices.
Related lessons
Keep going
Creators · 40 min
Reasoning Models (o-series, Claude Extended Thinking, Gemini Deep Think): When the Extra Tokens Are Worth It
When to spend 10x the tokens on a reasoning model — and when a normal model is fine.
Builders · 40 min
AI model families: DeepSeek and the China AI scene
Understand DeepSeek and why China's AI models surprised the world.
Creators · 10 min
The Reasoning-Model Family: When To Pay Extra For Thinking
The o-series, Opus thinking modes, Gemini Deep Think — reasoning models cost more per token but think before answering. Knowing when to pay is a money-and-time tradeoff.