Summary

  • OpenAI has released GPT-4.5, which the company claims is its best model for chat yet
  • The new release faces a tough crowd, with some critics contending that the company is ‘phoning it in’
  • The enhanced Large Language Model (LLM) is available to those with a $200-a-month ChatGPT Pro account, with a full release due next week
  • While the company has been tight-lipped about the specific enhancements, it has highlighted the fact that GPT-4.5 has fewer ‘hallucinations’ than its predecessor, GPT-4
  • Furthermore, the newest iteration performed better on SimpleQA, a general knowledge quiz, than GPT-4, with scores of 62.5% versus GPT-4’s 38.

By Will Douglas Heaven

Original Article