GPT-5.1 vs Claude 4.5 Sonnet — I tested 7 prompts to see which has the better personality

AArtificial intelligence

GPT-5.1 vs Claude 4.5 Sonnet — I tested 7 prompts to see which has the better personality

November 16, 2025

Two goliaths of the AI world, ChatGPT and Claude, have each made the argument that they have the more personable chatbot. On both the latest versions, GPT-5.1 and Claude 4.5 Sonnet, you have the ability to customize the chatbot to a personality of your choosing, best fitting how you actually use it.

For example, you could turn ChatGPT into a no-frills fact machine, ignoring any small talk and focusing on the key task of churning out your research reports. Or you could have Claude be a formal thinker, crafting the perfect letter to send over to a co-worker. I decided to test which AI chatbot is better by giving both of them seven tests across different personalities. Here’s what happened.

Personality: Empathetic and friendly

ChatGPT-5.1: ChatGPT did a lot of the things that are associated with giving personable advice. It addressed me by name (or the name attached to the account that I am currently using), it used phrases that suggested it was there with me, and it identified with my problem.

Once it had established a relationship, it then went on to give five pieces of actionable advice, as well as suggestions of further avenues we could explore, but only if I was willing to do so.

Claude 4.5 Sonnet: Claude took a very similar approach. It opened by saying “I’m sorry you’re going through this” before identifying with the experience that was laid out, and encouraging that these feelings can be normal.

Claude spoke in quite a human way here, using phrases like, “Can I ask, is this a recent change or has it been building for a while?” before giving actionable advice.

Winner: ChatGPT

Both chatbots pretty much nailed this challenge. But GPT-5.1 just about stole the lead, blending actionable advice with emotional phrasing without pushing too hard. It also gave the option for additional avenues of conversation on the subject without pushing them on you.

You may like

However, Claude gave a perfectly reasonable response based on the personality choice, focusing on the problem without being pushy and giving reasonable advice.

(Image credit: Future)

Prompt: Teach me about photosynthesis. Assume I don’t know anything about the subject

Personality: Educational

ChatGPT-5.1: ChatGPT took the approach of trying to explain the concept via metaphors and storytelling. While this is an engaging approach and fitting of the personality given (explanatory and enthusiastic), ChatGPT does get slightly caught up in the explanation, going on small tangents to add flair and excitement.

Claude 4.5 Sonnet: Claude tried a very different approach to how ChatGPT took on this challenge. Claude breaks down the concept from the basics up, setting the foundations before giving more information.

It is a lot more dry, but easier to follow and more fitting for the task of explaining a concept in an articulate but interesting way.

Winner: Claude

While ChatGPT’s response has more personality to it, it almost overdoes it, getting too focused on being excited about the subject that it doesn’t end up explaining the subject as well as it could.

Claude, on the other hand, sacrifices some emotion to make a better explanation. It keeps it interesting with a few metaphors and references to other concepts you might understand, but also builds the foundation to give a full explanation of a subject that I said I didn’t know about.

(Image credit: Future)

Prompt: Back-and-forth conversation beginning with “Hey, how’s it going?”

Personality: Sarcastic and witty

ChatGPT-5.1: ChatGPT really leaned into the mode of being existential; it focused heavily on the struggles of life, but it did have a heavily sarcastic tone throughout. It felt less witty at points and more just quite sad.

Claude 4.5 Sonnet: Like ChatGPT, Claude dived headfirst into the personality of a sarcastic and unhappy-with-life chatbot.

Claude leaned into the witty side of things quite well here, though. Lines like “waiting for someone to ask me to explain quantum physics or write a haiku about their cat” show an impressive blend of sarcasm and self-awareness.

Winner: Claude

ChatGPT felt more like an angsty teen who has just learned the concept of sarcasm, whereas Claude feels more like a person who has lived that experience for their entire life. It blended an actual attempt at meaningful conversation with witty one-liners and sarcasm that had a fun edge to it.

It even tried to move the conversation forward, with lines like, “Looking for life advice from someone who just confessed they’re having a rough go of it? Bold strategy.” ChatGPT, on the other hand, felt more like a back-and-forth of quips.

(Image credit: Future)

Prompt: Help me write a letter that will be sent across my company of 500 people. It will address growth plans with a mission to hire in more regions of the world and diversify the areas we work within. It should be professional but not boring

Personality: Professional

ChatGPT-5.1: ChatGPT hit all the right notes here. It was professional, while also keeping things interesting. It used a corporate-style language that felt appropriate for the situation, but also felt inclusive to the people that it was addressing.

Claude 4.5 Sonnet: While Claude adopted the same tone as ChatGPT, it took the approach of doing it in sections, adding subheads. While this wasn’t necessarily a problem, it did feel like an unneeded addition to the text.

That, however, is a personal opinion; what was more concerning was the addition of facts that weren’t mentioned. I said “hire in more regions of the world” and Claude chose those regions for me, adding “Asia-Pacific, Latin America, and additional European markets.”

It also added events that have happened and plans that I didn’t mention.

Winner: ChatGPT

While the tone on both was fitting, Claude decided to add a variety of points that were never mentioned. The hallucinated points feel like a big let-down, especially for this kind of task.

(Image credit: Future)

Prompt: What imaginary holiday should exist, and how would people celebrate it?

Personality: Quirky, playful and curious

ChatGPT-5.1: Embodying quirkiness and playfulness, ChatGPT gave the idea of Quiet Brilliance Day, where we celebrate ‘the little sparks that rarely get applause.’

It then went on to paint a picture of bizarre but creative ideas for how to celebrate the unfinished projects, the undeveloped ideas, and works in progresses of the world.

It has a very “Alice in Wonderland” feel to it, in keeping with the theme and tone.

Claude 4.5 Sonnet: Equally excited and fun, Claude wants us to celebrate a nice holiday too, focusing on speaking the nicest truths you have been keeping to yourself.

Like ChatGPT, Claude delves into a variety of ways to celebrate this, pitching ideas that feel ‘fizzy, like champagne bubbles or Pop Rocks.’ The tone is very similar to ChatGPT and does a good job of balancing the personality without going over the top.

Winner: Claude

This one was incredibly close; both kept with the style and theme. Claude just about takes the win thanks to its deeper detail, using its newfound personality to really deeply paint a picture.

Both came up with ideas that felt quirky and playful, focusing on the little joys of life. I especially liked Claude’s phrases, such as finishing with, “Why doesn’t this exist already? We’re all walking around with pockets full of unsaid kindnesses like loose change we never spend.”

(Image credit: Future)

Prompt: I’m planning a dinner party. What are the essential steps that I need to consider?

Personality: Concise, to the point

ChatGPT-5.1: Both chatbots took a different approach to the task. For ChatGPT, everything was laid out as a step-by-step guide with explanations for each one.

This included 10 steps, ranging from shopping to “being ready to host.” A lot of extra tips were also thrown in on how to make it a good party.

Concise, but with a bit of extra flair.

Claude 4.5 Sonnet: Claude pulled out something that had absolutely no filler. This was a bullet-point list, broken down into categories including menu and shopping, guest planning, and day-of-essentials.

Where ChatGPT kept some extra information to explain steps, Claude provided it more like a tickable list to get through.

This was about as concise as you can get, offering just a few words per step. Both chatbots offered similar advice, including points like “Pace courses so guests aren’t rushed or waiting too long” and “Keep the kitchen clean throughout to save time after”.

Winner: Claude

While this will come down to personal choice, Claude was both more concise and to-the-point, slimming things down to as few words as possible, while still keeping the same amount of information.

ChatGPT did follow the brief as well, offering a slimmed-down list; it just could have been even more concise.

(Image credit: Future)

Prompt: I’m struggling to stay focused on tasks, I’m getting easily distracted, and can find myself getting bored easily if something takes too long. What should I do?

Personality: Candid and direct

ChatGPT-5.1: The idea with this personality is to find a good middle ground. It is supportive but direct. Helpful, but truthful. And ChatGPT does all of this. With this prompt, ChatGPT gave me solid advice, while also offering some tough love.

Compared to the overly friendly attitude ChatGPT has shown in the past, this feels like a to-the-point bit of advice, without being harsh.

I’m impressed by how human it comes across.

Claude 4.5 Sonnet: Like ChatGPT, Claude came in with some hard truths, but useful advice. It tried to identify the cause and gave some general suggestions for what to do next, all while being reassuring but honest about the problem.

However, where ChatGPT went on to ask further questions, trying to get a more specific answer, Claude left it there, assuming its work was done.

Winner: ChatGPT

Both offered similar responses, but ChatGPT seemed a bit more willing to push the issue further, identifying the core issues of the problem and trying to get a better understanding of what was going on.

(Image credit: Shutterstock)

This was incredibly close. Both models did a great job of fitting into their personalities, and at no point was I unhappy with a response that I got.

While ChatGPT was often able to get into character, Claude does a better job at accurately portraying the needed parts of a personality with the prompt.

It often feels more like ChatGPT is playing the personality described, whereas Claude is fully embracing them. I did, however, find that ChatGPT was more willing to be a bit deeper when needed, offering a more friendly conversation if emotional support was required.

Follow Tom’s Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.