Particle.news
Download on the App Store
31 ARTICLES
·
3mo ago

OpenAI Unveils o3 and o4-mini AI Models with Visual Reasoning

The new models, available now for paid users, integrate image analysis and autonomous tool usage for advanced problem-solving.

OpenAI o3 and o4 have a parallel safety system that prevents these models from potentially handing out dangerous advice on biology and chemicals.
Image
ChatGPT
OpenAI cofounder and CEO Sam Altman. Today the company released two new AI "reasoning" models, o3 and o4-mini, as it seeks to show it can remain at the front of the AI pack.

Overview

  • OpenAI has launched its most advanced reasoning models, o3 and o4-mini, which are now available to ChatGPT Plus, Pro, and Team subscribers, as well as API users.
  • These models introduce visual reasoning capabilities, allowing them to analyze, manipulate, and reason with images, a first for OpenAI’s AI systems.
  • o3 and o4-mini can autonomously utilize all ChatGPT tools, such as web browsing, Python execution, and image generation, to solve complex, multi-step problems.
  • The o3 model delivers state-of-the-art performance in coding, math, science, and visual tasks, while o4-mini offers a faster, cost-efficient alternative with competitive capabilities.
  • OpenAI also introduced Codex CLI, a coding agent for local use, and announced a $1 million API-credit grant initiative to support early projects leveraging these models.