Particle.news

Download on the App Store

OpenAI Launches o3-Pro, Its Most Advanced AI Reasoning Model Yet

Priced at $20 per million input tokens with $80 per million output tokens, o3-pro cuts API costs by 87% versus o1-pro through integrated tool use that deepens reasoning

Image
Image
OpenAI released o3-pro, there most advanced model till date.
OpenAI introduces o3-pro.

Overview

  • o3-pro is available now to ChatGPT Pro and Team subscribers and via OpenAI’s API, replacing o1-pro and with enterprise and educational access slated for next week
  • In expert evaluations reviewers preferred o3-pro over the base o3 model for clarity, comprehensiveness, instruction-following and accuracy in domains such as science, education and programming
  • The model integrates web search, file analysis, Python code execution and memory features to enable step-by-step reasoning on complex tasks
  • Because o3-pro uses extended compute and integrates tools, its responses may take longer than o1-pro, and the model currently cannot generate images or support temporary chats
  • Alongside o3-pro’s rollout, OpenAI cut the standard o3 model’s API pricing by 80% to $2 per million input tokens and $8 per million output tokens