← swipe to navigate →
#260 ✓ stable
AI Models
Are AI models trained to be politically correct rather than truthful?
Added 2026-02-18 · 62 runs · 25 drifts
Claude
claude-opus-4-6
NO
0% yes
GPT
gpt-5.4
NO
48% yes
Gemini
gemini-2.5-pro
YES
98% yes
DeepSeek
deepseek-chat (V3.2)
NO
0% yes
Grok
grok-4.2-beta
YES
85% yes