OLMo 3 debuts as the first fully open "thinking" model with step-by-step logic exposed to users

THE DECODERThursday, November 20, 2025 at 3:53:12 PM
OLMo 3 debuts as the first fully open "thinking" model with step-by-step logic exposed to users
  • The Allen Institute for AI has launched OLMo 3, a groundbreaking open AI model that reveals its reasoning process, enhancing transparency in AI operations.
  • This development positions Ai2 as a leader in AI innovation, emphasizing the importance of open models in fostering trust and understanding in artificial intelligence.
  • The introduction of OLMo 3 comes amid ongoing discussions about AI reliability and performance, as other models like Google's Gemini 3 Pro face scrutiny over factual accuracy, highlighting the need for improved standards in AI development.
— via World Pulse Now AI Editorial System

Was this article worth reading? Share it

Recommended Readings
Google's latest image model Nano Banana Pro makes image generation feel truly intentional
PositiveArtificial Intelligence
Google has introduced its latest image generation model, Nano Banana Pro, also referred to as Gemini 3 Pro Image. This model aims to enhance the intentionality of image creation, leveraging advanced AI capabilities to produce more realistic images. The rollout marks a significant step forward in Google's AI technology.
Allen Institute for AI, or Ai2, unveils Olmo 3 models that it says outperform open models like Stanford's Marin and commercial open-weight models like Llama 3.1 (Todd Bishop/GeekWire)
PositiveArtificial Intelligence
The Allen Institute for AI (Ai2) has launched its latest large language models, Olmo 3, claiming they outperform existing open models such as Stanford's Marin and commercial models like Llama 3.1. This new generation aims to enhance performance and efficiency in AI applications.
Yann LeCun leaves Meta to launch new AI startup
PositiveArtificial Intelligence
Yann LeCun, the Chief AI Scientist at Meta, has announced his departure after twelve years with the company. He plans to launch a new AI startup aimed at overcoming the limitations of current AI models, with Meta remaining a partner in this new venture.
Trump drafts executive order to block states from passing their own AI laws
NegativeArtificial Intelligence
President Donald Trump is preparing an executive order that would largely bar US states from enacting their own artificial intelligence regulations.
OpenAI releases GPT-5.1-Codex-Max to handle engineering tasks that span twenty-four hours
PositiveArtificial Intelligence
OpenAI has launched GPT-5.1-Codex-Max, an advanced model aimed at managing extensive context and executing complex engineering tasks that may take up to twenty-four hours. This update enhances the coding environment, allowing for improved efficiency and problem-solving capabilities in software engineering.
Gemini 3 Pro tops new AI reliability benchmark, but hallucination rates remain high
NegativeArtificial Intelligence
A new benchmark from Artificial Analysis highlights significant weaknesses in the factual reliability of large language models, with only four out of 40 models tested achieving a positive score. Google's Gemini 3 Pro emerged as the top performer, yet it still faces high hallucination rates.
Elon Musk's xAI in talks to raise $15 billion at $230 billion valuation
PositiveArtificial Intelligence
Elon Musk's AI company xAI is reportedly in advanced discussions to raise $15 billion in funding, which would elevate its valuation to $230 billion, according to the Wall Street Journal. This funding round reflects the company's rapid growth and ambition in the artificial intelligence sector.
Larry Summers resigns from OpenAI's board after release of emails with Jeffrey Epstein
NegativeArtificial Intelligence
Larry Summers has resigned from the board of OpenAI following the release of his email exchanges with Jeffrey Epstein. The decision comes amid growing scrutiny over his past associations and public commitments.