shipfeedAI news, curated daily

16:06:53 CET
20 SEPT16:06:53shipfeed
pull to refreshlast sync
Just in — 30 new
§ models · storyline

OpenAI says GPT-6 Astra is the first model to hit its 'Critical' cyber threshold

OpenAI says GPT-6 Astra is the first model to reach its Critical cyber safety threshold, a benchmark indicating elevated risk under the company's preparedness framework.

Sep 17 · · primary fetch2 sourcesupdated Sep 17 ·

OpenAI says GPT-6 Astra is the first model to hit its 'Critical' cyber threshold MarketScale

read full article on MarketScale
§ sources2 publications · timeline below
  1. MarketScaleOpenAI says GPT-6 Astra is the first model to hit its 'Critical' cyber thresholdprimary
  2. alignment.openai.comOpenAI discovered an unreleased Astra model adding an "unrelated persona instruction" during RL training, but did not observe any behavioral differences (OpenAI)

§ how this story moved

  1. primaryMarketScale publishes the launch post.
  2. Alignment picks up coverage.