SafetyPage

Lab system cards and frameworks · System card · OpenAI

ChatGPT Images 2.0 System Card

21 Apr 2026

Read at deploymentsafety.openai.com ↗Read at sourceAll rights reserved

Outline · 52 paragraphs

  1. ChatGPT Images 2.0 System Card¶2
  2. 1. Introduction¶3
  3. 2. Observed Safety Challenges, Evaluations, and Mitigations¶6
  4. 2.1 New Safety Challenges¶7
  5. 2.2 Safety stack¶10
  6. 2.3 Safety and Policy Evaluations¶21
  7. 2.3.1 Automated Evaluations and Adversarial Testing¶22
  8. Table 1: Results¶33
  9. 2.3.2 Analysis¶37
  10. 2.4 Preparedness Framework: Image-Specific Capability Assessment and Safeguards¶41
  11. 2.4.1 Live Blocking¶44
  12. 2.4.2 Offline conversation review, flagging, blocking¶46
  13. 3. Image Provenance¶48