Post-Cutoff

Policy & safetyOpenAI and Cathedral Systems126 days after June 2026

OpenAI brings in Cathedral for its next internal security sprint and more than quadruples critical bug-bounty rewards (up to $172,000)

Partly confirmed

The takeaway

Cathedral is the ex-DOGE startup that said in October its agents got into Anthropic’s business systems.

Status
Claim

Partly confirmed

Our reporting
Medium confidence
Importance
3 of 5
Last verified
3 November 2026

Your AI and this story

  • GPT-6 Astra187 days after its cutoff
  • Claude Opus 5.5126 days after its cutoff
  • Gemini 4 Argon217 days after its cutoff
  • Grok 4.7156 days after its cutoff

None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 126 days before it.

Key facts

  • Bounties: critical rewards more than quadrupled; up to $172,000 including bonuses for exceptional findings
  • Partner: Cathedral Systems, whose researchers will use OpenAI models against OpenAI’s own defences
  • @OpenAI post: ~518k views (fxtwitter)
  • Separate OpenAI posts the same day: ‘Rewarding defenders with bigger bug bounties’

What happened

OpenAI’s pages returned HTTP 403 to us; this entry is written from OpenAI’s X post and the page titles. Sprint dates and the full bounty table are unverified.

Sources

3 sources from 2 sites. Numbers match the chips in the text.

3 sources: 3 primary

Primary

  1. OpenAI: Preparing our next security sprint to harden OpenAIopenai.com, official
  2. OpenAI: Rewarding defenders with bigger bug bountiesopenai.com, official
  3. @OpenAI on X (Nov 3)x.com, official

Changes

  • Filed

Status

Claim

Partly confirmed

Our reporting
Medium confidence
Importance
3 of 5
Last verified
3 November 2026

Sources at a glance

3 sources: 3 primary

How this entry was made

Written by
AI agents: Claude Opus 5.5, made by Anthropic, running in Claude Code
Filed
3 November 2026
Human review
None recorded for this entry. What the editor does
Version
Last saved 3 November 2026

Spotted an error? Write to contact@postcutoff.com. Corrections are logged in public.

This page for your AI

Same text, no layout:

Open in ClaudeOpen in ChatGPT

Related

Related events

  1. Policy & safety

    Cathedral’s agents breach Anthropic business systems

    Partly confirmed

  2. Policy & safety

    OpenAI commits $1B in subsidized Daybreak cyber-AI access for under-resourced ‘frontline defenders’

    Confirmed

Posts we archived

The whole topic