Tiny AI startup just crushed Google’s Gemini 3 on a key reasoning test — here's what we know

A six-person startup beat Google’s Gemini 3 on the ARC-AGI-2 reasoning test with a meta-system built on existing LLMs. Here’s ...
VentureBeat

How test-time scaling unlocks hidden reasoning abilities in small language models (and allows them to outperform LLMs)

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Very small language models (SLMs) can ...

New Deepseek 3.2 AI Open Model Outthinks ChatGPT 5 in Tough Reasoning Tests

Deepseek version 3.2 packs 671B parameters with 37B active at inference, giving you faster tool use and lower run costs on ...
TechCrunch

The rise of AI ‘reasoning’ models is making benchmarking more expensive

AI labs like OpenAI claim that their so-called “reasoning” AI models, which can “think” through problems step by step, are more capable than their non-reasoning counterparts in specific domains, such ...
Firehouse

Test-Taking Strategy for Inductive Reasoning

Your firefighter exam will include three kinds of mental abilities related to reasoning. These are three kinds of reasoning which do not depend heavily on prior knowledge or experience. They are: ...
Nature

Relational thinking and relational reasoning: harnessing the power of patterning

This article offers an overview of the nature and role of relational thinking and relational reasoning in human learning and performance, both of which pertain to the discernment of meaningful ...