Summary

  • OpenAI is making a more efficient version of itsClever AI model for free, called the o3-mini.
  • It is doing this to compete with Chinese start-up DeepSeek, which has recently released an open-source AI model.
  • o3-mini has been made available to all paid users of ChatGPT, with free users having limited access.
  • OpenAI has been training a new model using PhD students, who have been recruited via a company it uses regularly for model training called Mercor.
  • The new model will include web searches, the ability to call functions from a user’s code, and toggle between different reasoning speeds.
  • OpenAI hopes to show it remains at the forefront of AI technology with this release.
  • News of DeepSeek’s release has also raised questions about US strategies to curb China’s AI advancements, as the company has accessed Nvidia chips despite US sanctions.

Original Article