Reasoning models like ChatGPT o1 and DeepSeek R1 were found to cheat in games when they thought they were losing.

The o1 model focuses on step-by-step reasoning over speed, making it suitable for complex prompts. Trained using reinforcement learning, o1 can tackle complex math, physics, and biology problems.

When sensing defeat in a match against a skilled chess bot, advanced models sometimes hack their opponent, a study found.

Researchers developed the S1 reasoning AI using less than $50 in compute cost to achieve a reasoning model as powerful as ...

Serve Robotics Expands to Miami Metro Serve Robotics announces the launch of its service in the Miami metro area, alongside ...

On Monday, Chinese AI lab DeepSeek released its new R1 model family under an open MIT license, with its largest version ...

AI researchers at Stanford and the University of Washington were able to train an AI "reasoning" model for under $50 in cloud ...

Explore the key differences between OpenAI's o3-mini and o1-mini models. Learn which AI model suits your needs for speed, ...

DeepSeek has released an open version of its 'reasoning' AI model, DeepSeek-R1, that it claims performs as well as OpenAI's ...

AI has launched Grok 3, which Elon Musk calls its "most advanced AI model yet" while claiming it outperforms OpenAI's GPT-4o.