OpenAI Rolls Out GPT-6 Astra Amid Cyber Warnings — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI rolled out GPT-6 Astra to Plus, Pro, Enterprise, and Business Standard and Premium tiers in ChatGPT Work, Codex, and the API, with OpenAI calling it "the world's most intelligent and aligned model."
- GPT-6 Astra became OpenAI's first model to cross a "critical cybersecurity threshold," scoring 100% on ExploitBench, per InfoWorld and The Hacker News coverage.
- OpenAI released the model "with added cyber guardrails" per Bloomberg, while TechRadar reported OpenAI itself warned about how good Astra is at cracking cybersecurity before shipping it.
- Sam Altman publicly apologized for the "messy" GPT-6 Astra rollout that "locked out paying users," according to The Verge, with Unite.AI noting the staged launch initially denied paid access.
- Replit added GPT-6 Astra support one day after OpenAI's release, the fastest third-party integration called out in coverage.
- GPT-6 Astra introduces "human-level computer control" per iPhone in Canada, with The Tech Portal writing it "can do anything your computer can, and more."
- ChatGPT was reportedly down ahead of the Astra model launch per BleepingComputer, coinciding with the rollout.
Why it matters: GPT-6 Astra is being shipped simultaneously as OpenAI's most powerful model and its first to score 100% on ExploitBench, forcing paying customers to evaluate unprecedented capability against explicitly flagged cyber risk. The fact that Altman had to apologize for a rollout that "locked out paying users" shows OpenAI couldn't align access tiers, safety guardrails, and demand for a frontier release.
Ask SkimNews
