Geo News
Community curated by people like you
LatestAICryptoHealthWorld AffairsUS Politics

GPT-5 stories

Apr 16, 2026

OpenAI Releases GPT-5.4-Cyber Model for Defensive Cybersecurity with Restricted Access

OpenAI launched GPT-5.4-Cyber, a model specifically trained for defensive cybersecurity applications, with access initially restricted to verified security experts. The release follows Anthropic's Mythos model and represents OpenAI's strategic response in the cybersecurity AI space.

Apr 16, 2026·3 sources
00

Top claims

  • ▪Anthropic announced Claude Mythos Preview is only being privately released because the company says it could be exploited by hackers and bad actors.
  • ▪OpenAI released a $10 million cybersecurity grant fund in February 2026.
  • ▪Claude Mythos was released approximately one week before GPT-5.4-Cyber.

Topics

AI securityLarge language models (LLMs)AI foundation models

Related entities

GPT-5.4

Featured Stories

OpenAI pauses development of Astra AI model over critical cybersecurity capability concerns

OpenAI pauses development of Astra AI model over critical cybersecurity capability concerns

Aug 7, 2026·2 sources