AINewsnow

GPT-6 Astra Tops ExploitBench With Perfect Security Score

This story is from 2026-09-05. It is preserved in the archive; the latest stories are on the live feed.

Forensic Summary OpenAI's GPT-6 Astra achieves 100% on ExploitBench and 99.2% on binary reverse engineering benchmarks, significantly outperforming its predecessor GPT-5.6 Sol on security-relevant tasks. The model's exceptional capability at offensive security benchmarks raises dual-use concerns, a…

Read the full story at DEV Community — AI ↗

Timeline · 1 report

  1. 2026-09-05 14:31 · DEV Community — AI
    GPT-6 Astra Tops ExploitBench With Perfect Security Score

More stories

  1. Introducing Astra for Law — OpenAI News
  2. OpenAI discloses six new safety incidents — Axios AI+
  3. Microsoft exec called AI scraping the “largest theft of labor in human history” — Ars Technica AI
  4. Sources: Anthropic expects to generate $100B+ in annualized revenue this year, up from $65B as of July, as it moves ahead with its IPO amid the AI safety debate (New York Times) — Techmeme
  5. OpenAI staff knew the ‘existential threat’ AI posed to publishers, New York Times claims — Financial Times AI
  6. How Cooley is accelerating IPO work with ChatGPT — OpenAI News
  7. Helping older adults use AI in everyday life — OpenAI News
  8. How to connect AI usage to business value — OpenAI News

Get the daily brief of stories like this at 6:30 every morning →