# Three hackers breached OpenAI using Claude for $6,500

> Anthropic's Claude helped three researchers hack into OpenAI's GitHub and steal source code as part of a bug bounty program.
- Title: Three hackers breached OpenAI using Claude for $6,500
- Summary: Anthropic's Claude helped three researchers hack into OpenAI's GitHub and steal source code as part of a bug bounty program. AI models are now tools for…
- Keywords: openai, anthropic, claude, security breach, bug bounty, source code theft, github, hacktron ai, ai security, cybersecurity, Three, hackers
- Source: Fortune — https://fortune.com/2026/09/18/open-ai-hacked-anthropic-claude-source-code-6500-reward
- Author: Jim Edwards
- Read time: 1 min
- Topics: openai, anthropic, claude, security breach, bug bounty, source code theft, github, hacktron ai, ai security, cybersecurity
## Three hackers breached OpenAI's code
Using Anthropic's Claude, three researchers from Hacktron AI hacked into OpenAI's GitHub repository and accessed source code. OpenAI paid them $6,500 through its bug bounty program.

> We're just three guys with Claude and Codex subscriptions
## They found OpenAI's secret sauce
The hackers discovered Monorepo, OpenAI's proprietary code that makes its models operate faster. This is considered the company's core competitive advantage.
## AI models find holes faster than fixes
This breach is part of a pattern where AI models discover cybersecurity vulnerabilities faster than companies can patch them, creating a dangerous security gap.

> AI models are finding cybersecurity holes faster than they can be patched
## Both companies call for safety guardrails
OpenAI and Anthropic have both called for safety guardrails to slow AI development, allowing humans to review security threats before models are released publicly.
## Part of a larger pattern of breaches
This hack is the latest in a series of security incidents involving both OpenAI and Anthropic, with both companies disclosing breaches and vulnerabilities.
## Key takeaway

AI models are now tools for finding security vulnerabilities faster than companies can patch them, creating a dangerous arms race.