Who makes OpenAI hack via libheif (Hacktron)?

OpenAI hack via libheif (Hacktron) is made by Hacktron.

Does OpenAI hack via libheif (Hacktron) work in the EU?

Unverified. No explicit evidence found. This is a security research disclosure, not a product with regional availability. Last verified 14 Nov 2026.

How much does OpenAI hack via libheif (Hacktron) cost?

Pricing unknown. Not a product. Researchers reportedly spent under $3,000 in tokens; OpenAI reportedly paid a $6,500 bounty.

Where was OpenAI hack via libheif (Hacktron) shown?

In 1 creator video by AI Search. Most recently by AI Search on 20 Sept 2026, at 31:56 in "OpenAI hacked, Jev, Google’s RSI, Qwen 3.8 Omni, Bonsai 2, new Gemini Live: AI NEWS".

OpenAI hack via libheif (Hacktron)

by Hacktron

Security research report on an AI-assisted attack chain against OpenAI systems using a libheif bug and an SSO flaw.

UnverifiedResearch Agents & AutomationResearch

Updated 14 Nov 2026EU status checked 14 Nov 2026Last shown 20 Sept 2026

OpenAI hack via libheif (Hacktron) at a glance

What is OpenAI hack via libheif (Hacktron)?: OpenAI hack via libheif (Hacktron) is a research from Hacktron. Security research report on an AI-assisted attack chain against OpenAI systems using a libheif bug and an SSO flaw.
Who makes OpenAI hack via libheif (Hacktron)?: OpenAI hack via libheif (Hacktron) is made by Hacktron.
Does OpenAI hack via libheif (Hacktron) work in the EU?: Unverified. No explicit evidence found. This is a security research disclosure, not a product with regional availability. Last verified 14 Nov 2026.
How much does OpenAI hack via libheif (Hacktron) cost?: Pricing unknown. Not a product. Researchers reportedly spent under $3,000 in tokens; OpenAI reportedly paid a $6,500 bounty.
Where was OpenAI hack via libheif (Hacktron) shown?: In 1 creator video by AI Search. Most recently by AI Search on 20 Sept 2026, at 31:56 in “OpenAI hacked, Jev, Google’s RSI, Qwen 3.8 Omni, Bonsai 2, new Gemini Live: AI NEWS”.

Sources: creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.

About

This entry covers a security research disclosure by Hacktron, as described by a YouTube creator (AI Search). According to that account, three Hacktron researchers gained access to OpenAI's internal systems in under 72 hours, using Claude to assist the work. No primary source was available for verification, so details below rely on the creator's summary.

The reported chain combined a bug in libheif, an image-decoding library, triggered through the OpenAI community forum, with a flaw in single sign-on (SSO). Together these were said to allow takeover of ChatGPT and Codex accounts. The cost in model tokens was reported as under $3,000.

OpenAI reportedly patched the issues and paid a $6,500 bug bounty. This is a research disclosure, not a product or model: there is nothing to install, access or license, and no official page, code or paper was available to confirm the technical details.

Key features

Reported libheif image-decoder bug reached via the OpenAI community forum
Chained with an SSO flaw to take over ChatGPT and Codex accounts
Reportedly done in under 72 hours with Claude assisting
Reported token cost under $3,000
Reportedly patched by OpenAI with a $6,500 bounty

1 moment

As seen on YouTube

Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.

AI Search20 Sept 2026

OpenAI hacked, Jev, Google’s RSI, Qwen 3.8 Omni, Bonsai 2, new Gemini Live: AI NEWS

Three Hacktron researchers broke into OpenAI's internal systems in under 72 hours using Claude. They chained a libheif image-decoder bug on the community forum with an SSO flaw to take over ChatGPT and Codex accounts, for under $3,000 in tokens. OpenAI patched it and paid a $6,500 bounty.

31:56Open on YouTube