o1-preview / o1-mini

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

o1-preview / o1-mini — Learning to Reason with LLMs

Why it matters

The first commercial reasoning model; on AIME 2024, from 12% for GPT-4o to 74% for o1

Source — Primary sources first; secondary reports are labelled “Press”

Officialopenai.comhttps://openai.com/index/learning-to-reason-with-llms/

Added in Week 35, 2026 · See that week

On the map — ⑥ Reasoning Models line

  1. Previous stopClaude 3.5 Sonnet
  2. This record · stopo1-preview
  3. Next stopGoogle introduces Gemini 2.0
See it on the map

In the model family tree — GPT family · 23 versions

  1. Earlier versionThe first version in this branch
  2. This recordo1-preview / o1-mini
  3. Later versiono3-mini
See the whole GPT family

Same week

Upstage releases Solar Pro PreviewOfficialAnt Group releases KAG (Knowledge Augmented Generation)Paper

Nearby on the same axis

xAI releases Grok-2 / Grok-2 mini in betaOfficialSakana AI announces The AI Scientist, with Oxford and UBCOfficialScaling LLM Test-Time Compute OptimallyPaperGoogle DeepMind announces AlphaProof and AlphaGeometry 2 results at IMO 2024Official