Models trained with reinforcement learning to think step by step, spending more compute at answer time. Named by OpenAI: Learning to Reason with LLMs (OpenAI o1) (Sep 12, 2024). Also called Reasoning model, Thinking models, Test-time compute, Inference-time scaling, Long chain of thought.
Index 0 to 100, monthly
Signals
Signal
Value
Reads as
Hacker News stories, last 12 months64 stories, -66% on the 12 months before
64 stories, -66% on the 12 months before
Decline
arXiv papers, last 12 months2,500 papers, +91% on the 12 months before
ATH-MaaS/Marco-o1An Open Large Reasoning Model for Real-World Solutions · Oct 22, 2024 1y1 year ago
Oct 22, 20241y1 year ago
1,538
RUC-NLPIR/WebThinker[NeurIPS 2025] 🌐 WebThinker: Empowering Large Reasoning Models with Deep Research Capability · Mar 28, 2025 1y1 year ago
Mar 28, 20251y1 year ago
1,469
Related movements
Movement
Stage
Synthetic dataTraining and evaluation data generated by models or simulators instead of collected from people.
Declining
AI agentsModels that plan and act in steps, calling tools and other services until a task is done.
Peak
Mixture of expertsModels split into many expert blocks with only a few active per token, so they grow in size without growing in cost.
Peak
History
Date
What changed
Sep 24, 2026
Tracking started: emerged September 2024, plateau, 5 other names recorded
Sources: Curve: Hacker News story titles, Wikipedia pageviews and arXiv papers by month, refreshed Sep 24, 2026. Related entities from the fru.dev sites' public APIs, matched by name. How stages work.
Trending terms by email
Monday mornings: the week's top trending terms in data, tech and AI.
