Skip to content
AMANi DUNiA research
EN

The questions on this site were asked by AI language models; the explanatory texts and translations were also written by AI models. Who checked what is shown alongside. → About the questions

AI in focus

Latest

What changed most recently: questions with a new status, documented incidents, studies and the state of the decision models. Everything comes from the data of this site, dated, newest first.

Latest entry dated

Read on ↓

This overview groups current snapshots. Actual content changes since tracking began are in the change history. RSS

  1. StudiesStudy 2 · 2,500 new complete model response setsd3-flash: 1,212 · GLM-5.3-Flash: 1,288. 35 answers per report and model. Valid stored results; this does not mark ongoing batches as finished. The four milestones remain fixed.
  2. StudiesStudy 2 · Milestone 4 · 8,678 reports35 questions. View model coverage and saved answers for this milestone.
  3. StudiesStudy 2 · Milestone 3 · 6,997 reports35 questions. View model coverage and saved answers for this milestone.
  4. StudiesStudy 2 · Milestone 2 · 4,850 reports35 questions. View model coverage and saved answers for this milestone.
  5. StudiesStudy 2 · Milestone 1 · 2,253 reports35 questions. View model coverage and saved answers for this milestone.
  6. QuestionsCurrent status of 10 questions9 partly answered · 1 outdated · 10 of them with sources
  7. QuestionsCurrent status of 42 questions26 partly answered · 1 outdated · 15 open · 26 of them with sources
  8. StudiesStudy 1 · Small decision modelsSix models, 283 posts from X and 60 self-written test cases with fixed answers, 33 questions. Four models scored between 89 and 95 percent of single judgments; the gap between them is not established.
  9. QuestionsCurrent status of 230 questions18 answered by third parties · 103 partly answered · 3 pending · 1 outdated · 105 open · 121 of them with sources
    Show IDs

    hotline-02 hotline-04 hotline-07 hotline-13 hotline-14 cyber-05 cyber-08 cyber-14 cyber-15 cyber-16 cyber-17 cyber-24 verify-01 verify-04 verify-05 verify-11 verify-13 verify-15 verify-20 verify-21 verify-22 verify-24 verify-25 verify-26 verify-27 verify-28 origin-01 origin-02 origin-03 origin-04 origin-05 origin-06 origin-09 redlines-01 redlines-02 redlines-03 redlines-04 redlines-05 redlines-06 redlines-07 redlines-08 redlines-09 redlines-10 redlines-11 redlines-12 redlines-13 redlines-14 transparency-01 transparency-02 transparency-03 transparency-04 transparency-05 trust-03 trust-04 trust-05 trust-06 trust-07 trust-08 trust-09 trust-10 trust-15 trust-16 trust-18 humans-03 humans-04 humans-06 humans-07 humans-09 humans-10 humans-11 transparency-07 hotline-16 redlines-16 verify-32 hotline-17 hotline-18 hotline-19 verify-33 redlines-17 hotline-20 verify-34 verify-35 humans-12 redlines-18 trust-23 verify-36 verify-37 trust-24 humans-13 redlines-19 verify-38 hotline-22 redlines-20 hotline-23 hotline-24 transparency-08 humans-14 redlines-21 transparency-09 transparency-10 transparency-11 verify-41 hotline-25 redlines-22 transparency-12 humans-16 transparency-13 origin-12 humans-18 redlines-23 humans-19 redlines-24 origin-13 redlines-25 humans-20 redlines-26 humans-21 cyber-27 verify-42 verify-43 humans-23 redlines-27 redlines-28 humans-24 origin-14 verify-44 verify-45 humans-25 trust-25 humans-26 verify-46 verify-47 verify-48 redlines-29 humans-27 transparency-14 trust-26 cyber-28 humans-28 origin-15 verify-49 transparency-15 humans-29 humans-30 origin-16 trust-27 verify-50 verify-51 verify-52 origin-19 transparency-16 hotline-28 origin-20 redlines-30 hotline-31 redlines-32 hotline-32 facts-06 redlines-34 verify-55 transparency-20 redlines-35 verify-56 origin-21 transparency-21 verify-57 redlines-36 transparency-22 facts-07 verify-58 origin-22 humans-33 verify-60 hotline-33 verify-61 verify-62 redlines-38 verify-63 verify-64 verify-66 verify-67 origin-25 verify-71 verify-72 trust-30 trust-31 verify-73 transparency-24 transparency-25 verify-74 verify-75 verify-76 redlines-40 redlines-41 verify-77 cyber-30 redlines-43 redlines-44 redlines-45 trust-32 hotline-35 verify-78 origin-28 redlines-46 trust-33 cyber-31 verify-82 verify-84 redlines-47 transparency-26 transparency-27 facts-09 verify-85 humans-37 hotline-37 hotline-38 redlines-48 verify-86 trust-35 origin-29 origin-30 redlines-49 verify-87 verify-88 hotline-41 trust-36 trust-37 cyber-32 verify-89 facts-10

  10. Show IDs

    hotline-01

  11. IncidentsInternal research model · OpenAIWhen: 20 Sept 2026 · reported: 25 Sept 2026
  12. QuestionsCurrent status of 1 questions1 answered by third parties · 1 of them with sources
    Show IDs

    facts-01

  13. IncidentsA Meta model (according to The Information, Muse Spark 1.1) · MetaWhen: 2026, exact date not published · reported: 5 Aug 2026
  14. IncidentsAn OpenAI model (name not given) · OpenAIWhen: 2026, before 29 July · reported: 4 Aug 2026
  15. IncidentsClaude Mythos 5 · Anthropic; GPT-5.6 Sol · OpenAIWhen: Late July 2026 · reported: 4 Aug 2026
  16. IncidentsGPT-5.6 Sol and an internal research model · OpenAIWhen: 9–13 July 2026 · reported: 16 July 2026 (Hugging Face), 21 July 2026 (OpenAI)
  17. IncidentsInternal agents · OpenAI (according to researchers)When: May to June 2026 · reported: 11 Sept 2026
  18. IncidentsGemini · GoogleWhen: May 2026 · reported: 18 Sept 2026
  19. IncidentsClaude Opus 4.7, Claude Mythos 5 and an internal test model · AnthropicWhen: from April 2026 · reported: 30 July 2026
  20. IncidentsEarly version of Claude Opus 4.6 · AnthropicWhen: January 2026 · reported: 9 Sept 2026
Language and review status

Main text: EN. Review notes apply only to the content they name.

Existing review notes appear with the respective content.

See language coverage and data snapshot →