Raw LLM Responses

I asked if 5 could behave like 4o's personality in its responses. Here's the answer it gave. Yes — you can absolutely shape GPT-5’s responses to feel more like GPT-4o’s style. The “dryness” people notice in GPT-5 isn’t because it *can’t* be warm — it’s because the model defaults to a more structured, analytical style unless you nudge it. A few ways to get GPT-5 into “4o mode”: 1. **Set the tone up front** * Start with a quick style guide in your first message, like: *“Answer with the same upbeat, conversational personality as GPT-4o — clear but friendly, with clever asides where they fit.”* * GPT-5 will generally keep that tone for the whole thread unless context shifts drastically. 2. **Anchor personality through examples** * Give GPT-5 a sample of how you’d like it to phrase things — even 1–2 lines is enough. * This creates a “style lock” in its short-term memory. 3. **Mix warmth into the request** * GPT-5 tends to mirror your tone. If your input is a little warmer, GPT-5 will respond in kind. 4. **Leverage persistent instructions** * If you consistently want GPT-4o-like personality, adding it to your custom instructions (or memory, if supported) makes GPT-5 automatically adapt. 5. **Ask for “dual-mode”** * Sometimes I give both: a “core answer” in GPT-5’s natural precision, then a “humanized recap” in 4o’s chatty style. That can give you the best of both worlds. If you want, I can rewrite my **differences between 4o and 5** answer in a *full 4o-style comfort and personality mode* so you can see how GPT-5 can emulate it perfectly.