o1

Definition

Plain language

OpenAI's first reasoning model trained to think out loud before answering.

Also called: o3, o3-mini

Why it matters: It demonstrated that scaling test-time reasoning, not just model size, can dramatically improve performance on hard problems.

For example, o1 might spend tens of seconds writing internal reasoning before producing the final answer to a competition math problem.

Heard on the show

“Take o3-mini, OpenAI's reasoning-specialized model, and GPT-4o-mini, its ordinary sibling.”

Mentioned in 8 episodes

1. 197
   Twin Problems Suggest AI Reasoning Gains Are Mostly Better Fact Recall
2. 140
   When a Reasoning Model Says "Let Me Double-Check" After It's Already Decided
3. 124
   A Cheap Model With the Blueprints Beats Expensive Models Working Blind
4. 082
   Training a Deep Research Agent on 8,000 Synthetic Tasks: The Rubric Tree Trick
5. 053
   An AI Agent Swapped In Focal Loss And Beat A Human-Tuned Training Script
6. 041
   When the Iteration Teaches the Model to Skip the Iteration
7. 028
   Teaching a Model to Hire Copies of Itself: Recursive Agent Optimization
8. 026
   What RL Actually Does to Language Models, at the Token Level

Related terms