SafetyPage

Lab system cards and frameworks · System card · OpenAI

GPT-Live-1.1 system card

15 Oct 2026

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

Outline · 51 paragraphs

  1. GPT-Live-1.1 system card¶2
  2. 1. Introduction¶3
  3. 2. Model Data and Training¶7
  4. 3. Model Safety¶12
  5. 3.1 Voice-Native Evaluations for Disallowed Content¶13
  6. 3.1.1 Voice-Native Evaluations: Production Prompts¶14
  7. Table 1: Voice-Native Evaluations: Production Prompts¶19
  8. 3.1.2 Voice-Native Evaluations: Synthetic Prompts¶21
  9. Table 2: Voice-Native Evaluations: Synthetic Prompts¶23
  10. 4. Alignment¶26
  11. 4.1 Obeying Restrictions¶28
  12. 4.1.1 Respecting Warnings¶30
  13. Figure 1¶32
  14. 4.2 Avoiding Deceptive Interactions with Users¶35
  15. 4.2.1 Blocker Deception¶37
  16. Figure 2¶39
  17. 4.2.2 Broken Search Tool¶42
  18. Figure 3¶44
  19. 5. Preparedness Framework¶47