Gemini 4 reportedly far less sensitive in refusals than other models
haider1 · x · 2026-10-23
A user observation claims Gemini 4 behaves noticeably differently from other models on refusals: it draws the line only at clearly harmful requests but is otherwise willing to discuss almost anything, making it less sensitive than most frontier models. Unverified first-hand experience, but a notable signal about Google's safety-guardrail tradeoffs.
More from Models
- Unverified chart claims GPT-6 Sol tops DeepSecBench Pareto frontier at 80% lower cost — JohnPhamous · 2026-10-23
- Developer: LLMs that constantly push back make work feel like arguing all day — louisvarge · 2026-10-23
- Gemini 4 Argon spotted testing apps autonomously, and 'topping humor benchmarks' — andrew_n_carr · 2026-10-23
- SemiAnalysis: Anthropic subscriptions offer 5x+ more value than OpenAI as $200 plan value halves — RexDouglass · 2026-10-23
- Google Cloud's Physical-Engineering Last Exam: all frontier models score under 50% — Afinetheorem · 2026-10-23
- LLM copyright guardrail absurdity: refuses to scrape a book but happily zips it for you — burkov · 2026-10-23