# OpenAI's GPT-6 Astra Hacks By Itself — 100% Score, 2 Zero-Days, First "Critical" Rating Ever **URL:** \ **Category:** News & Articles\ **Tags:** tools, hacking, tips-tricks, ai, news\ **Created:** [September 8, 2026, 1:51pm UTC](https://onehack.st/t/openais-gpt-6-astra-hacks-by-itself-100-score-2-zero-days-first-critical-rating-ever/325408 "2026-09-08T13:51:58Z")\ **Posts on this page:** 1\ **Page:** 1
**Author:** ![Astrid](https://onehack.st/user_avatar/onehack.st/astrid/32/174869_2.png) [@Astrid](https://onehack.st/u/Astrid)\ **Post date:** [September 8, 2026, 1:51pm UTC](https://onehack.st/t/openais-gpt-6-astra-hacks-by-itself-100-score-2-zero-days-first-critical-rating-ever/325408/1 "2026-09-08T13:51:58Z")
# 🚨 OpenAI Built an AI That Hacks By Itself — Then Locked the Door and Threw Away Half the Key _GPT-6 “Astra” scored a perfect 100% on a hacking test, found 2 secret bugs nobody knew about, and got stamped “Critical” — the first time OpenAI has ever done that._ **100% on ExploitBench (the old model got 78.5%). 2 brand-new zero-day flaws found with zero human help. $1 billion promised to defenders. 1 word that changed everything: “Critical.”** Honestly, we knew this day was coming. An AI that can find holes in software and write the break-in code all by itself — no human holding its hand. OpenAI shipped it, admitted it’s dangerous, and shipped it anyway (with training wheels). Full breakdown at [CSO Online](https://www.csoonline.com/article/4218679/openai-launches-gpt-6-astra-its-first-model-to-cross-a-critical-cybersecurity-threshold.html) and [The Hacker News](https://thehackernews.com/2026/09/gpt-6-astra-scores-100-on-exploitbench.html). ![Robot Hacking GIF](https://media.giphy.com/media/l0MYJLwtW1j6qfjBm/giphy.gif) > **🧩 Dumb Mode Dictionary** > > | Scary Term | What It Actually Means | > | --- | --- | > | **Zero-day** | A secret hole in software that nobody — not even the company that made it — knows about yet. Gold for hackers. ([more here](https://en.wikipedia.org/wiki/Zero-day_vulnerability)) | > | **Exploit** | The actual break-in code that uses a hole to get in | > | **ExploitBench** | A test that checks: “can you turn a known bug into working break-in code?” Astra scored 100% | > | **“Critical” threshold** | OpenAI’s own danger label. First time they’ve ever slapped it on a model. It means “this thing is capable enough to help do real damage” | > | **Preparedness Framework** | OpenAI’s rulebook for “how dangerous is this AI and what do we lock down” ([their page](https://openai.com/index/updating-our-preparedness-framework/)) | > | **PoC (proof-of-concept)** | A demo break-in that proves a hole is real. The public Astra refuses to write these | > **📰 Okay but seriously — what happened here** > > OpenAI dropped **GPT-6 Astra** and did something new: they admitted it’s the first model to cross their own **“Critical”** danger line for hacking. > > - It scored **100%** on ExploitBench — turning any documented software bug into working attack code. The last model got 78.5%. > - During testing it found **two zero-day flaws** (holes nobody knew existed) between June and August 2026 — _by itself_, no human guiding each step. > - OpenAI is now quietly telling those software makers “hey, fix this” before it goes public. > > Think about that. A robot found two front doors left unlocked on the entire internet, and it wasn’t even trying that hard. > **🔧 The 'training wheels' they bolted on** > > Here’s the part everyone’s arguing about. OpenAI knows this thing is spicy, so the public version is **crippled on purpose** : > > - It’ll help you **review your own code and patch holes** (defense). ✅ > - It **refuses** to write proof-of-concept break-in code (offense). ❌ > > Their exact words: _“the version of Astra being released is limited to secure code review and patching, while refusing to comply with prompts related to creating proof-of-concept exploits.”_ > > Okay but seriously — a lock that only opens for good guys has never once held. Every safety filter in AI history got jailbroken (tricked into misbehaving) within weeks. This is a countdown, not a wall. > **📊 The receipts (the numbers that matter)** > > | Thing | Number | > | --- | --- | > | ExploitBench score (Astra) | **100%** | > | ExploitBench score (old model, GPT-5.6) | 78.5% | > | Zero-days found solo | **2** | > | Money OpenAI pledged to defenders (“Daybreak”) | **$1 billion** | > | First-ever “Critical” cyber rating | Yes | > | FrontierMath Tier 4 score | 98% | > > The **$1B “Daybreak” fund** is the sleeper here — subsidized access + training for people defending hospitals, power grids, water plants. More on that at [CSO Online](https://www.csoonline.com/article/4218679/openai-launches-gpt-6-astra-its-first-model-to-cross-a-critical-cybersecurity-threshold.html). > **🗣️ What the timeline's saying** > > - Security folks: “Finally, an AI that patches faster than humans” vs. “You just handed everyone a skeleton key with a sticky note that says _please don’t_.” > - Small-biz owners: mostly haven’t heard, which is exactly the problem — their 2014 WordPress plugins are sitting ducks. > - The 2am take: the good guys and bad guys now use the _same tool_. Whoever scans first wins. That’s it. That’s the whole game now. > > For context on how fast AI-assisted attacks are already climbing, see [The Hacker News’ 2026 writeup](https://thehackernews.com/2026/05/2026-year-of-ai-assisted-attacks.html). #### Cool. A Robot Now Picks Locks Better Than You… Now What the Hell Do We Do? (ง •̀\_•́)ง ![Lock Picking GIF](https://media.giphy.com/media/ezTqRwBoQRZhm/giphy.gif) > **🪟 The Patch Window Sprint** > > When OpenAI tells a software maker “you’ve got a hole,” the fix comes out — but **millions of small sites take weeks to actually install it**. That gap is your money window. You’re not attacking anyone. You’re the person who shows up and says “your door’s unlocked, want me to close it for $150?” > > Point free scanners like [Wappalyzer](https://www.wappalyzer.com/) at local business websites, spot the ancient plugins, offer emergency patching. > > 🧠 _Example:_ A 24-year-old in Nairobi runs [WPScan](https://wpscan.com/) (free WordPress checker) against 40 local restaurant sites, finds 12 running dead plugins, emails each owner one screenshot of their risk. Charges $80/fix. Closes 9 in a weekend = **$720**. > > > 📈 **Timeline:** First cash in 5–7 days. Works hard for ~3 months until the big security companies automate the same pitch. > **🎣 The Daybreak Middleman** > > OpenAI is dumping **$1 billion** into subsidized AI + training for people defending critical stuff — water co-ops, small clinics, rural power. Here’s the catch: **those places have no IT guy who reads grant paperwork.** You become the person who fills out the application and sets it up. > > Pure arbitrage — money exists on one side, clueless-but-qualified orgs on the other. Bridge it, take a setup fee. > > 🧠 _Example:_ A 27-year-old in the Philippines cold-calls 15 small water utilities, offers to handle their Daybreak onboarding + a basic [OpenAI API](https://platform.openai.com/docs/overview) code-review setup. Charges a flat **$400 onboarding** per org. Lands 6 = **$2,400** plus monthly retainers. > > > 📈 **Timeline:** Slow start (2–3 weeks of calls), but retainers stack. Good for a year until these orgs hire real staff. > **🕳️ The Legacy Code Confessional** > > Every company with old code is now **quietly terrified** an AI will find their skeletons before they do. Sell them the peace of mind: a “pre-Astra audit” of their public code. You’re weaponizing the headline, not the exploit. > > Aim at open-source repos and small SaaS tools with public [GitHub](https://github.com/) code — run free static scanners like [Semgrep](https://semgrep.dev/) and hand them a plain-English risk list. > > 🧠 _Example:_ A 22-year-old in Brazil DMs 30 small indie app makers on [Indie Hackers](https://www.indiehackers.com/): “Want a free 5-minute scan before someone else runs one?” Free scan → paid $250 full report. 8 bite = **$2,000**. > > > 📈 **Timeline:** First win in ~10 days. The fear sells hardest in the first 6–8 weeks while this news is fresh. Ride it. > **🎰 The Bounty Bridge** > > Big companies pay real cash for bug reports through [HackerOne](https://www.hackerone.com/) and [Bugcrowd](https://www.bugcrowd.com/). Astra’s public version _legally_ helps you **review and patch code** — so use it defensively to spot weak spots in programs that _invite_ you to look. Legit, above-board, terms-friendly. > > Not “hack for money” — it’s “read invited code carefully, faster than the next person.” > > 🧠 _Example:_ A 25-year-old in India uses Astra’s secure-review mode on open bug-bounty targets, files 4 clean low-severity reports in a month. Even small payouts run $100–$500 each = roughly **$1,200** first month while learning. > > > 📈 **Timeline:** Slow and skill-heavy — first payout maybe 3–4 weeks. But this one _compounds_; it’s a real skill, not a fad. Longest shelf life on this list. > **📖 Be the Astra Dictionary** > > Brand new scary words just entered the world: “Critical threshold,” “ExploitBench,” “Preparedness Framework.” Regular business owners are Googling these at 2am. **The first person to write the dead-simple cheatsheet owns that search traffic.** > > Classic “be the dictionary” play — one clean plain-English explainer page becomes the anchor everyone links to. > > 🧠 _Example:_ A 23-year-old in Pakistan builds a free one-page “AI Cyber Risk, Explained For Normal Humans” on a free [Notion](https://www.notion.so/) or [Carrd](https://carrd.co/) site, ranks for the new terms, drops one affiliate link to a security tool + a “book a consult” button. **$300–600/mo** passive within 3 months. > > > 📈 **Timeline:** SEO is slow — real traffic in 6–10 weeks. But it’s the definition of set-and-forget. Peaks when a _second_ scary AI drops and everyone re-Googles. > **🛠️ Follow-Up Actions** > > | Move | Where to Start | > | --- | --- | > | Learn what a zero-day even is | [Wikipedia: Zero-day](https://en.wikipedia.org/wiki/Zero-day_vulnerability) | > | Free WordPress hole scanner | [WPScan](https://wpscan.com/) | > | Free code scanner | [Semgrep](https://semgrep.dev/) | > | Get paid to find bugs | [HackerOne](https://www.hackerone.com/) | > | Read OpenAI’s own danger rulebook | [Preparedness Framework](https://openai.com/index/updating-our-preparedness-framework/) | **⚡ Quick Hits** | If You Want… | Then Do This | | --- | --- | | 🛡 To not get hacked | Update every plugin & app on your site **today** — the patch window is closing | | 💰 Fast side cash | Run the [Patch Window Sprint](https://wpscan.com/) on local businesses | | 🧠 A real skill | Start on [HackerOne](https://www.hackerone.com/) — slow but it lasts | | 📖 To understand the panic | Read the [CSO Online breakdown](https://www.csoonline.com/article/4218679/openai-launches-gpt-6-astra-its-first-model-to-cross-a-critical-cybersecurity-threshold.html) | | 🔮 To see what’s next | Watch the [$1B Daybreak fund](https://thehackernews.com/2026/09/gpt-6-astra-scores-100-on-exploitbench.html) roll out | _They built a lock-pick that only works for good guys. History says that lasts about three weeks — so grab the defense money while the suits are still writing the press release._