Google, Anthropic, and OpenAI have released new AI cybersecurity models and safety initiatives to aid defense while addressing risks of model misuse and unauthorized actions.
Reader brief
No Reader read of this story yet.
What was said3
Attributed, verbatim. Every quote is checked against the article it came from. One that does not match is not shown.
Tulsee Doshi
1 quote“With Gemini 3.8 Flash Cyber, we focused specifically on equipping defenders with expert capabilities that give them an advantage over attackers. This is why we have invested in vulnerability fixing from the start, and prioritized it over offensive capabilities like exploitation”
Anthropic
1 quote“It refused malicious agentic coding and computer use requests at a comparable rate to Mythos 5, Sonnet 5, and Opus 5, and it is our most robust model to date on an external prompt injection benchmark.”
OpenAI
1 quote“Based on that work, we believe Astra's safeguards sufficiently minimize the risk of severe harm for release under our Preparedness Framework.”
Sources1
All filed from IndiaNamed United States · CrowdStrike · Datadog · Hugging Face · Menlo Security · Palo Alto Networks · Snowflake · Anthropic · Astra · Claude Fable 5.1 · Claude Mythos 5.1 · Gemini 3.8 Flash Cyber · Google
- [1]The Hacker NewsneutralGoogle, Anthropic, and OpenAI Unveil Cyber AI Models, Safeguards, and Access Programs