Skip to content

Search stories, models, companies…Subscribe

Topic · Concept

Reasoning models

Reasoning models are language models trained to work through a problem in intermediate steps, sometimes called "thinking," before giving a final answer. The approach improves results on math, coding and multi-step tasks, but the extra tokens add cost and latency.

Follow via RSS Get Reasoning models alerts3 stories · updated Oct 9

Fact file

Chain-of-thought promptingWei et al., Jan 28, 2022

OpenAI o1 / "Learning to reason with LLMs"September 12, 2024; RL on chain of thought; scales with test-time compute

DeepSeek-R1Jan 22, 2025; reasoning via reinforcement learning

Claude thinking controlsbudget_tokens (extended thinking): Claude 4.5 and earlier; deprecated on 4.6; rejected (400 error) on 4.7 and later, incl. the 5.5 models and Fable 5.1, which use adaptive thinking plus effort. Thinking counts toward max_tokens and billed output tokens

Facts verified Oct 9

Start here

ExplainerReasoning models explained: thinking tokens, what they cost, and when to use them

Latest on Reasoning models

Oct 9Models · News6 minGPT-6 reaches every ChatGPT user: what OpenAI shipped, what it costs, and where each model runs Oct 9Research · Analysis6 minOpenAI's Navier–Stokes claim, explained for builders: the verifier is the story Oct 9Models · Explainer7 minReasoning models explained: thinking tokens, what they cost, and when to use them

Nothing in this format on this page — try another filter or the next page.

Timeline

SEP 2024OpenAI o1 / "Learning to reason with LLMs"