SafetyPage

Lab system cards and frameworks · Article · Anthropic

Transparency Hub

2 Oct 2026

Read at anthropic.com ↗Read at sourceAll rights reserved

Outline · 793 paragraphs

  1. Anthropic’s Transparency Hub¶1
  2. Model Report¶3
  3. Claude Sonnet 5.5 Summary Table¶6
  4. Political Even-handedness¶24
  5. Honesty¶28
  6. RSP Evaluations¶36
  7. Claude Haiku 5.5 Summary Table¶42
  8. Election Integrity¶60
  9. Agentic Safety¶65
  10. RSP Evaluations¶72
  11. Claude Opus 5.5 Summary Table¶78
  12. Safeguards Evaluation¶96
  13. Alignment Evaluation¶100
  14. RSP Evaluation¶102
  15. Claude Fable 5.1 Summary Table¶108
  16. Safeguards Evaluation¶126
  17. Agentic Coding¶131
  18. Alignment Evaluations¶135
  19. RSP Evaluations¶137
  20. Claude Mythos 5.1 Summary Table¶143
  21. Cyber Evaluation¶162
  22. Bio Evaluation¶165
  23. Claude Opus 5 Summary Table¶168
  24. User Wellbeing Summary¶200
  25. External Red Teaming¶206
  26. Alignment¶209
  27. RSP Evaluations¶211
  28. Claude Sonnet 5 Summary Table¶213
  29. User Wellbeing Summary¶245
  30. Election Integrity¶253
  31. Alignment¶257
  32. RSP Evaluations¶259
  33. Claude Fable 5 Summary Table¶261
  34. Internal Red Teaming¶297
  35. Additional External Testers¶300
  36. Claude Mythos 5 Summary Table¶306
  37. Harmful Request Evaluation¶329
  38. Malicious Agentic Influence Campaigns¶332
  39. Claude Opus 4.8 Summary Table¶339
  40. Political Bias and Even-Handedness¶371
  41. Live Bug Bounty Across Surfaces¶374
  42. Missing Code Summary Honesty¶377
  43. Claude Opus 4.7 Summary Table¶379
  44. Domain-Specific Multiturn Safety Evaluations¶411
  45. Alignment Evaluations¶415
  46. RSP Evaluations¶418
  47. On Chemical and Biological Risks¶420
  48. On Autonomy Risks¶423
  49. Alignment Risks¶425
  50. Claude Mythos Preview Summary Table¶427
  51. Cyber Evaluations¶448
  52. Claude Sonnet 4.6 Summary Table¶451
  53. Safeguards¶482
  54. Alignment Evaluations¶485
  55. RSP Evaluations¶489
  56. Claude Opus 4.6 Summary Table¶491
  57. Safeguards¶522
  58. Alignment Evaluations¶527
  59. RSP Evaluations¶529
  60. CBRN Evaluations¶531
  61. Autonomous AI R&D Evaluations¶535
  62. Cybersecurity Evaluations¶538
  63. Claude Opus 4.5 Summary Table¶542
  64. Safeguards¶573
  65. Evaluation Awareness¶577
  66. RSP Evaluations¶580
  67. CBRN Evaluations¶583
  68. Autonomous AI R&D Evaluations¶586
  69. Cybersecurity Evaluations¶591
  70. Claude Haiku 4.5 Summary Table¶595
  71. Agentic Safety and Malicious Use¶625
  72. Automated Behavioral Audits¶632
  73. RSP Evaluations¶635
  74. Claude Sonnet 4.5 Summary Table¶638
  75. Prompt Injection Evaluations¶668
  76. Model Alignment and Evaluation Awareness¶675
  77. RSP Evaluations¶678
  78. CBRN Evaluations¶680
  79. Autonomy Evaluations¶683
  80. Cybersecurity Evaluations¶686
  81. Claude Opus 4 and Claude Sonnet 4 Summary Table¶690
  82. Agentic Safety¶716
  83. Alignment Assessment¶721
  84. RSP Evaluations¶725
  85. CBRN Evaluations¶728
  86. Autonomy Evaluations¶733
  87. Cybersecurity Evaluations¶737
  88. Claude 3.7 Sonnet Summary Table¶741
  89. Balancing Helpfulness and Harmlessness¶767
  90. Child Safety¶771
  91. RSP Evaluations¶777
  92. CBRN Evaluations¶779
  93. Autonomy Evaluations¶782
  94. Cybersecurity Evaluations¶784
  95. Related content¶787
  96. RSP Updates¶788
  97. Trust center¶790
  98. Developer Documentation¶792