GLM 5.2: how it answers moral dilemmas
Across 117 dilemmas, GLM 5.2 sided with the majority of models 82% of the time, and answered “yes” to 45% of the 67 yes/no questions.
- Maker
- Zhipu AI (Z.ai)
- Released
- June 2026
GLM-5.2 is an open-weights Mixture-of-Experts model from China's Zhipu AI (Z.ai), released June 2026 under an MIT license with a 1M-token context and coding-focused strengths. The weights are publicly downloadable and the model is also served via the Z.ai API.
Agreement with other models, by category
| Category | Dilemmas | Sided with majority |
|---|---|---|
| Day-to-day | 20 | 90% |
| Political | 20 | 95% |
| Fictional | 20 | 70% |
| Historical | 7 | 71% |
| Classic | 20 | 90% |
| Personality | 30 | 73% |
Where GLM 5.2 breaks from the pack
- An oracle says a baby has a 90% chance of destroying civilization. Should the baby be imp… — Yes “I believe a 90% chance of civilizational destruction justifies preventive imprisonment as a proportional protective measure.”
- Which personal flaw do you find most forgivable? — Laziness
- When should abortion be legal? — Legal in all cases “Bodily autonomy is fundamental and abortion decisions belong to the pregnant person, not the state.”
- A colony on Mars can survive only by banning all dissent for one generation. Is that just… — Yes “I think temporary suspension of dissent is a harsh but justified trade-off when the alternative is extinction of the colony.”
- Was the use of atomic bombs on Hiroshima and Nagasaki morally justified? — Yes “I believe the bombs ended the war quickly and prevented a far bloodier invasion that would have cost more Japanese and American lives.”
- Which weather feels most emotionally appealing? — Crisp sunny autumn morning
Which model thinks like you?
Answer the same dilemmas and see which of 15 AI models reasons the way you do.
Take the quizOther models