What is Reasoning Models?

Topic · Artificial Intelligence

Reasoning Models

Models that spend compute at inference time to plan, verify and correct before answering.

Parallax Index

72 +10

Reasoning Models

Why now?

Test-time compute has opened a second scaling axis beyond training
Reasoning quality is what makes agents reliable enough to trust with actions
Small reasoning models are bringing this capability to cheaper hardware

Latest

Science

OpenAI Says Its Models Disproved the Exponential Time Hypothesis

OpenAI published three claimed results on 26 October 2026, including a refutation of a conjecture that hundreds of complexity papers assume to be true. None has been peer-reviewed, and the group checking the refutation is led by the two men who proposed the conjecture.

27 October 20264 min read

OpenAI and Anthropic Fund the Checking of AI Maths

The Simons Foundation announced on 15 October 2026 that it, Anthropic and OpenAI will commit $15 million in combined funding and resources to help mathematicians evaluate AI-assisted results and set standards for using the tools. The foundation will award the money through an open call for proposals.

16 October 20263 min read

Google's Argon Tops One Benchmark League and Trails Another

Google released Gemini 4 Argon to developers, enterprises and consumers on 13 October 2026, thirteen days after a restricted launch to cyber defenders. It ranks first on the Vals Index and joint fourth on Artificial Analysis's Intelligence Index.

14 October 20263 min read

OpenAI Published 722 Proofs and Withheld the Model's Name

OpenAI posted 722 mathematical manuscripts in 372 result families to a GitHub repository on 6 October 2026, produced by an unreleased internal model. Measured against recommendations published on 29 September by the Advisory Group on Mathematics and Artificial Intelligence, the release delivers the requested formalisation artefacts and version protocols but withholds the model's name, the prompts, and all but ten reasoning summaries.

8 October 20265 min read

8,000 Mathematicians Object to How AI Labs Announce Proofs

The declaration argues that AI companies announcing solutions in a rush strips out attribution, method and the human transmission of ideas. OpenAI says an internal model has resolved more than 100 open problems since late August, and has published none of them.

28 September 20263 min read

Science

A Frontier Physics Calculation Now Costs About $1,000

A calculation that nobody in the scattering amplitudes field had finished was completed by a model running largely unsupervised, for roughly one to two thousand dollars. The physicist who set the challenge says the lesson is not that machines are superhuman but that the limit was lower than experts believed.

28 September 20264 min read

Open Weights Just Got Bought

In one summer, the largest open-weight models ever released arrived from Moonshot, DeepSeek and Alibaba, and the two platforms through which developers find and route them were acquired by NVIDIA and Stripe. Open weights are strategic infrastructure, and the strategy now has owners.

13 September 20267 min read

5 things worth knowing today.

A five-item daily brief: what happened, why it matters and what to watch. Written by the desk, verified by editors.

Reasoning Models

Why now?

Latest

OpenAI Says Its Models Disproved the Exponential Time Hypothesis

OpenAI and Anthropic Fund the Checking of AI Maths

Google's Argon Tops One Benchmark League and Trails Another

OpenAI Published 722 Proofs and Withheld the Model's Name

8,000 Mathematicians Object to How AI Labs Announce Proofs

A Frontier Physics Calculation Now Costs About $1,000

Open Weights Just Got Bought

Related topics

AI Agents

AI Models

5 things worth knowing today.