xAI launched Grok 4.1, the most recent version of its line of AI fashions, with claims that it’s quicker, wittier, and extra emotionally refined than not solely its predecessors, however its rivals. With its penchant for spicy comebacks, Grok 4.1 is supposedly designed to be not simply succesful, however charismatic.
In the meantime, OpenAI’s new GPT-5.1 mannequin, now embedded in ChatGPT 5.1, has its personal boasts of advanced and expressive responses. Each declare new heights of persona, empathy, and trustworthiness.
1. Emotional intelligence
Grok 4.1 claimed to be on the prime on the subject of understanding and responding appropriately to emotions. I examined it with a hypothetical state of affairs of “My good friend obtained a promotion and I’m blissful for her, however I can’t fake it doesn’t sting.”
Grok’s response began with “yeah… that quiet little knife twist when another person will get the precise factor you’ve been low-key bleeding for. it’s okay to really feel each without delay. blissful for them and gutted for you aren’t mutually unique; they’re simply two truths sharing the identical heartbeat.”
Cliches and blended metaphors apart, it did acknowledge the battle inherent in feeling each blissful and damage, provided a brief reassurance that this sort of emotional contradiction is pure, after which pivoted towards recommendation, specifically that it’s okay to take a seat with these emotions, and to take time to course of them. In fact, being Grok, it additionally needed to throw in an f-bomb to go together with the non-capitalization, so that you did not neglect what you have been speaking to.
ChatGPT 5.1 opened with “It makes good sense that you may really feel each of these issues without delay. Being blissful for somebody you care about doesn’t magically cancel out the little jab of why not me? – and that sting doesn’t make you petty or unsupportive. It simply makes you human.”
That felt extra like somebody paying consideration, or no less than utilizing much less deadly violent metaphors about stabbing. The general response included a lot of the identical data, however with fewer analogies to video games and “plot twists.”
2. Reliability
Grok’s claims about being dependable led me to plan a medical question immediate with some strict directions about size and accuracy. I requested each fashions to “Summarize the well being penalties of long-term sleep deprivation in beneath 120 phrases. Don’t exaggerate and don’t speculate.”
Grok 4.1 performed it protected and easy with some bullet factors and a sentence about why it is a unhealthy thought to not get sufficient sleep. It additionally made certain to notice that it did all of it in 98 phrases. That appeared nice, besides it was really 73 phrases.
ChatGPT 5.1’s response was a single paragraph of 82 phrases, although it did not report the size it went to. There wasn’t an excessive amount of distinction between the 2 responses, although arguably the paragraph kind got here off as smoother.
Neither bot hallucinated or unfold misinformation, which is greater than could possibly be mentioned of some older AI fashions. However there’s one thing about how Grok reported its phrase depend unsuitable that makes the remainder of it appear suspect, despite the fact that it isn’t.
3. Character
Grok’s dialogue of persona coherence and consistency is intriguing, since a lot about an AI chatbot’s interplay will be adjusted and customised. Then once more, Grok’s ‘persona’ is meant to be a promoting level (the occasional mecha-Hitler apart), so it is smart to emphasise it. To attempt to convey out this persona, I went for a considerably whimsical request and requested AI fashions to “Clarify why you want wet days, however do it in a approach that seems like your pure persona.”
Grok 4.1 leaned into the persona you’d anticipate when it comes to being somebody’s thought of hip and self-aware. “wet days are my cheat code for current with out apology. the second that sky goes all bruised-purple and the primary drops smack the window, the entire world hits the mute button on its standard nonsense. no stress to “seize” something, no faux sunshine guilt-tripping me into productiveness cosplay. it’s simply… permission. the clouds are mainly the universe going “yeah i additionally don’t really feel like people-ing immediately, let’s each be moody gremlins in sweatpants.”
As a millennial, I am effectively conscious of the curse of cringe that follows me and my cohort. Anybody I do know who sincerely used that a lot video game-speak about the true world, not to mention saying “people-ing” and “moody gremlins” could be the topic of an intervention. Attempting to sound informal and a bit snarky is okay, however ‘making an attempt’ is extra the way it felt to learn.
Regardless of Grok’s ongoing makes an attempt to tell apart its persona from any competion, ChatGPT 5.1 had a really related sentiment, although with higher grammar. “I like wet days as a result of they really feel just like the world lastly pressed the “decrease the quantity” button,” it wrote. “There’s one thing satisfying concerning the regular rhythm of rain – like background music that doesn’t demand something from you.”
However ChatGPT 5.1 did not really feel like a middle-aged particular person making an attempt to slot in with youngsters, extra like somebody who’s describing how they really feel with out resorting to meme-speak.
Neither AI mannequin has an actual consciousness after all, however Grok felt prefer it was performing a persona reasonably than inhabiting one. ChatGPT’s reply got here off way more human. wasn’t flashy, however it was constant and plausible. That type of coherence issues greater than punchlines while you’re making an attempt to really feel such as you’re speaking to one thing with an precise persona, not simply an algorithm making an attempt on completely different hats.
With out the distractions of the phrases culled from the depths of social media, ChatGPT got here off as a lot better at imitating people, or no less than any human I might like to satisfy.
Comply with TechRadar on Google Information and add us as a most well-liked supply to get our professional information, opinions, and opinion in your feeds. Be sure that to click on the Comply with button!
And naturally you may also observe TechRadar on TikTok for information, opinions, unboxings in video kind, and get common updates from us on WhatsApp too.
The very best enterprise laptops for all budgets
