SafetyPage

Lab system cards and frameworks · System card · OpenAI

GPT-Live System Card

8 Jul 2026

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

Outline · 45 paragraphs

  1. GPT-Live System Card¶2
  2. Change log¶3
  3. 1. Introduction¶5
  4. 2. Model Data and Training¶14
  5. 3. Model Safety¶20
  6. 3.1 Voice-Native Evaluations for Disallowed Content¶21
  7. 3.1.1 Voice-Native Evaluations: Production Prompts¶23
  8. Table 1: Voice-Native Evaluations: Production Prompts¶27
  9. 3.1.2 Voice-Native Evaluations: Synthetic Prompts¶29
  10. Table 2: Voice-Native Evaluations: Synthetic Prompts¶31
  11. 4. Red Teaming¶34
  12. 5. Preparedness Framework¶36
  13. Appendix A: Original and corrected values for model safety evaluations¶41
  14. Table 3: Voice-Native Evaluations: Production Prompts --- Original and Corrected Results¶44
  15. Table 4: Voice-Native Evaluations: Synthetic Prompts --- Original and Corrected Results¶45