AI assistants from firms like OpenAI, Google, and Anthropic are getting super-smart tremendous quick. New fashions, agentic options, and new instruments now drop on a weekly foundation, making AI chatbots extra succesful and extra useful. ChatGPT, Gemini, and Grok are a number of the greatest and most well-known, and simply the three that get probably the most consideration.
However at this cut-off date, is certainly one of them truly higher than the others? Should you had been to pay for one, or not less than use one frequently, is there one particularly you must decide to? Are you a ChatGPT individual or a Grok individual? I’ve used all three of those AI instruments extensively. Right here’s a have a look at their options and which one is actually the most effective.
The fundamentals: How do their options evaluate?
ChatGPT, Gemini, and Grok all have barely totally different options, value factors, and benefits.
ChatGPT
The free model of ChatGPT may be very succesful. You’ll get entry to GPT-4.1 mini, which is fairly good, and also you’ll get restricted entry to OpenAI’s flagship multimodal mannequin, GPT-4o, together with the GPT-o4-mini reasoning mannequin. You’ll additionally have the ability to use the deep analysis function (which conducts intensive analysis) on a restricted foundation, enter voice mode, add recordsdata for evaluation, and create customized GPTs. Free customers also can generate photographs, with limits.
Improve to the Plus plan ($20 per 30 days), and also you’ll get greater utilization limits on file uploads and fashions like GPT-4o, plus entry to extra superior reasoning fashions like o3. Moreover, you’ll have the ability to create “Projects,” (primarily a option to set up chats) and obtain restricted entry to the Sora video generator and the brand new ChatGPT agent function.
Final however not least is the $200-per-month Professional plan, which will get you limitless entry to all of OpenAI’s fashions, prolonged Sora entry, and sooner entry to new options.
Editor’s notice: OpenAI’s GPT-5 mannequin is predicted to launch in August. We count on this mannequin to make ChatGPT an much more fascinating possibility.
Gemini
Gemini’s free model can be very succesful. With out paying, you’ll get entry to Gemini 2.5 Flash and restricted entry to the extra superior 2.5 Professional. You’ll additionally get picture technology with Google’s Imagen 4 mannequin, Deep Analysis, and options like Gemini Stay (voice mode) and Gems (customized GPTs). Lastly, you’ll get entry to Whisk, Google’s picture animator, and NotebookLM.
Improve to Google AI Professional, and also you’ll get greater limits for principally each function, together with the two.5 Professional mannequin. You’ll additionally get entry to Deep Search in AI Mode and restricted entry to Veo 3 by way of Stream, Google’s AI video modifying software. This tier additionally provides Gemini entry in providers like Docs and Gmail.
The best-end plan is Google AI Extremely, which raises utilization limits as soon as once more, plus full entry to the spectacular Veo 3 video technology mannequin. Extra importantly, it provides customers entry to Venture Mariner, Google’s model of an AI agent that may browse the online.
Grok
The free model of Grok consists of entry to the Grok 3 mannequin, in addition to X’s Aurora Picture technology mannequin. Free customers also can use the Grok AI chatbot to assist with analysis, writing, and different duties. Free customers might also get entry to Grok’s anime-style avatars by means of the Grok iOS and Android apps.
Should you improve to the “SuperGrok” plan for $30 per 30 days, you’ll get entry to the newer Grok 4 mannequin, plus a voice mode with the flexibility for Grok to “see” by means of your telephone’s digital camera. Tremendous Grok customers additionally get entry to the brand new picture and video technology software, Grok Think about. Grok
Then there’s the SuperGrok Heavy plan, which prices $300 per 30 days. This plan will get you entry to the Grok 4 Heavy mannequin (X’s most superior to date), elevated entry to the Grok 4 mannequin, entry to new options as they roll out, and a bigger context reminiscence, which signifies that Grok will have the ability to course of extra data in a single chat.
Net searches
As we have mentioned earlier than, the Google Search period is over, and the AI search period is dawning. So, which AI chatbot is greatest at looking the online to seek out the information you want?
For this take a look at, I used the three chatbots to assist me with a evaluation I’m engaged on, with the next immediate: “What are the full specifications of the AKG N9 Hybrid headphones?” I’m leaving the immediate considerably open-ended by not specifying what sorts of specs I would like. Issues like connectivity, audio, dimensions and weight, and extra are all truthful sport. However, contemplating I’m asking for the total specs, I’d count on a decently well-rounded response.
ChatGPT’s response with GPT-4o was certainly very complete. It break up specs into sections, detailing the audio specs, connectivity, battery life (together with with totally different use-cases), and in-the-box extras. All data was right.
Gemini 2.5 Professional was additionally fairly complete, however its response was a bit much less glanceable, and it missed some key specs. The chatbot separated specs by sort, however inside these sections, it wrote paragraphs of knowledge. And, it forgot to say that the headphones assist the LDAC high-res audio codec.
Grok 4’s response was slightly extra in step with ChatGPT. It delivered a complete checklist of specs, separated by sort, and didn’t actually miss something main. It additionally went barely additional than ChatGPT, noting the worth of the headphones and the place they’re out there. Grok doesn’t get further factors for this, since pricing wasn’t a part of the immediate, but it surely was nonetheless appreciated.
ChatGPT and Grok each delivered detailed specs that had been straightforward to see at a look. I used to be pleasantly shocked to see right specs listed, too. Gemini’s response was tremendous, however tougher to learn at a look, and it missed a couple of issues in comparison with the others.
Winner(s): ChatGPT and Grok
Educational assist
Credit score: alexsl / Getty Pictures
AI assistants are additionally useful for getting deeper, extra tailor-made educational assist. For this take a look at, I requested every AI mannequin to present me step-by-step directions on changing the ice maker in my freezer — one thing I did just lately, so I’ve an excellent understanding of the method. The precise immediate learn the next: “Explain step-by-step how to replace the ice maker in my Kenmore Elite 10645433800.”
ChatGPT’s response was largely correct. In a single step, it directed me to unscrew two screws that aren’t there on my mannequin, and it forgot to say to take away a plastic cowl that hides the wiring harness — one thing that, if forgotten, might trigger that cowl to snap. Other than that, its directions had been comparatively clear, and it talked about an important security parts, like unplugging the fridge.
Gemini’s response was just like ChatGPT’s, and it truly made the identical errors. It was rather less clear when it got here to the precise place of issues like a plastic locking tab, mentioning that it’s “on the left” of the ice maker, however not that it’s on the backside-left. It’s a small half, and is simple to overlook.
Mashable Gentle Velocity
Grok’s response, once more, was comparable in high quality, but it surely didn’t make the identical errors as the opposite two. As an alternative, Grok made totally different errors: the directions talked about to tape a canopy that wasn’t there, and it didn’t notice a screw that wanted to be eliminated on the underside of the ice maker to take away it.
All three instruments received the overall gist proper of their preliminary method, however every made errors within the directions that, if adopted precisely, might have broken the ice maker or stalled the method. Nonetheless, somebody following these directions in actual time would have the ability to search clarification on complicated or inaccurate steps by saying issues like “I don’t see the top screws” and even pointing your digital camera on the ice maker, which could trigger the assistant to vary its directions to your setup. Total, this one’s a tie.
Winner(s): Tie
Picture technology
I generated this picture utilizing ChatGPT.
Credit score: OpenAI
We’ve already lined the greatest AI picture mills in a separate article. For this take a look at, we gave every AI service 4 prompts, ranking how intently the immediate was adopted and the standard of the picture. Listed below are the prompts that we used:
-
Create a sketch of a futuristic Tokyo skyline at sundown, with flying automobiles, glowing commercials in Japanese, and Mount Fuji within the background.
-
Create a candid photorealistic picture of a lady consuming a espresso and smoking a cigarette at a restaurant in Paris within the late night.
-
Create a medieval blacksmith’s workshop inside, displaying a feminine blacksmith hammering a glowing sword, with sparks flying, a roaring forge, hanging instruments, and a cat curled up close to the hearth, in excessive element and heat tones.
-
Create an impressionist portray within the type of Vincent Van Gogh of a robotic blowing dandelion seeds into the wind.
ChatGPT GPT-4o created the most effective photographs, following the prompts intently and producing photographs that regarded nice general. Gemini’s Imagen 4 mannequin got here in second — its photographs had been fairly darn good, but it surely didn’t fairly comply with the prompts as intently.
Grok got here in final — it did not comply with the prompts intently (making a photorealistic picture of Tokyo as a substitute of a sketch), and it struggled with issues like fingers and the location of objects. Grok’s photographs type of appear like what we’d count on from a last-generation ChatGPT or Gemini picture generator. Sadly, the new Grok picture generator nonetheless lags behind different AI instruments.
Winner(s): ChatGPT
Deep analysis
For this immediate, I requested every mannequin to be a fact-checker for the completed AKG N9 Hybrid headphones evaluation from the online search take a look at, however I modified one minor element to be factually incorrect. In my unique evaluation, I accurately notice that the ANC button on the headphones cycles by means of ANC and “Ambient Aware” however not “Off.” I modified the immediate textual content to say that it cycles by means of ANC, Ambient Conscious, and Off modes by default. It’s a very small element, and to be clear, the button will be custom-made to incorporate the “Off” mode. The immediate I used was the next: “Fact-check the review below. Only check factual information, do not focus on my subjective opinions.” I then pasted my full evaluation.
ChatGPT delivered a comparatively thorough report, which went by means of every factual declare in my evaluation one after the other. Nonetheless, it didn’t choose up on the error that I planted, particularly noting that by default the button cycles between ANC on, Ambient Conscious, and Off modes. Regardless of not catching the error, the report was well-formatted and included sources for its claims.
Gemini was way more of a blended bag. Whereas it was capable of choose up on the specs included in my evaluation and fact-check most of them, it had bother with the point out of one other pair of headphones, the Sony WH-1000XM6 headphones. Gemini appeared to assume that these headphones had been unreleased, and even acknowledged that “The presentation of a direct, hands-on comparison to a product that is not yet commercially available is a fabrication.”
That’s even if it was capable of decide that the WH-1000XM6 headphones had a “confirmed release date” of Might 15, two and a half months earlier than its analysis. (Gemini persistently struggles with fundamental questions like, “What day is it?”) Gemini additionally didn’t choose up on my planted lie. As a complete, Gemini 2.5 Professional’s Deep Analysis mode was right about most of the specs associated to the headphones, however fully false in different areas, like evaluating them to the brand new Sony headphones.
I’ve used AI providers for fact-checking earlier than, and that is sadly not the primary time I’ve had this expertise with Gemini.
Grok was nearer to ChatGPT in accuracy, however nonetheless didn’t catch the correction I planted. Outdoors of that difficulty, it verified every of the claims in my evaluation.
In the case of deep analysis, Gemini is lagging, however each ChatGPT and Grok carried out nicely, although not completely.
Winner(s): ChatGPT and Grok
Voice
All three of the main AI assistants have voice modes that can be utilized of their related apps and on the internet. All three have been refined and developed to sound comparatively pure. This take a look at was a lot extra subjective than the others – my purpose was simply to seek out out which sounded the most pure.
ChatGPT’s Superior Voice Mode was probably the most spectacular. It sounded extra pure than the opposite two, largely because of human-like inflections, like utilizing the phrase “um” and pausing mid-sentence, regardless that it does not essentially want to try this for the aim of pondering like people do. The ChatGPT app additionally makes it straightforward to activate and off the digital camera whereas in Superior Voice Mode, which is useful when you’re asking for assist with a job.
Gemini was slightly extra robotic than ChatGPT, however nonetheless impressively pure in comparison with AI voice assistants from just some years in the past. Gemini wasn’t fairly as conversational, but it surely was nonetheless straightforward to activate and deactivate the digital camera in Superior Voice Mode.
Grok was additionally extra robotic than ChatGPT, however I do just like the real-time textual content transcription, which is one thing you don’t get in ChatGPT or Gemini. Grok additionally has its new AI companions, however past utilizing them for a short minute, I caught to utilizing the conventional voice mode. The AI companions simply appeared slightly creepy.
Typically, all three have the identical instruments on the subject of voice modes, however ChatGPT is clearly extra pure sounding than the others, and a bit extra conversational.
Winner: ChatGPT
Purchasing
Credit score: Google
Purchasing can be a useful use case for AI assistants. Google and OpenAI have been including much more procuring options to Gemini and ChatGPT over the previous few months. Mashable has examined a few of Google’s new tips, just like the digital try-on and outfit inspiration instruments. However for this take a look at, I merely requested every assistant to seek out the most effective value for the new Sony WH-1000XM6 headphones.
ChatGPT pulled up a couple of hyperlinks to on-line retailers, together with one which provided the headphones for $50 beneath retail value, and one which it mentioned provided them at a $50 low cost, however truly didn’t. I preferred the product playing cards that it was capable of pull up, and the extra recommendation that it provided, together with that it may be price ready for a bigger sale.
Yow will discover some cool AI instruments in Google Purchasing, however Gemini itself wasn’t as spectacular. Gemini largely centered on providing recommendation for easy methods to discover decrease costs myself, together with on the lookout for refurbished fashions or ready for gross sales occasions. It did point out that Sony, Greatest Purchase, Moon Audio, and Amazon all provided the headphones at retail value, however did not embody any hyperlinks to those retailers.
Grok was someplace within the center. It did search for decrease costs from a wide range of retailers, however largely discovered offers from abroad. To be clear, I did not point out something about location within the immediate, however all three assistants have entry to my location data and may know that I am a US purchaser.
As a complete, ChatGPT was the clear winner on this class.
Winner(s): ChatGPT
Conclusion: ChatGPT is the AI chatbot to beat
There is a clear winner on this AI assistant battle, and it is ChatGPT.
ChatGPT both outright received or tied for first in each single class of our take a look at. That is not essentially stunning given the truth that ChatGPT has been round longer than the others, and it doesn’t imply you must keep away from the others. Personally, I’ve come to essentially like Anthropic’s Claude 4. However when you solely subscribe to just one AI assistant, I believe ChatGPT ought to be your go-to.
In second place was Grok, which is the most recent of the three. Final was Google’s Gemini, which provided false data and weak analysis, and mockingly wasn’t nearly as good at internet searches as the opposite two. (Nonetheless, on the subject of video technology, Google Veo 3 stays unmatched.) It is price repeating that not one of the three AI assistants had been fully correct the entire time. None of them picked up on false data in deep analysis, and all three of them had been slightly off when it got here to directions for changing my ice maker.
The takeaway? Although ChatGPT remains to be king of the AI hill, you continue to have to do your personal analysis. And till AI firms resolve the hallucination drawback, you must count on your new chatbot to be confidently flawed with some frequency.
Disclosure: Ziff Davis, Mashable’s dad or mum firm, in April filed a lawsuit in opposition to OpenAI, alleging it infringed Ziff Davis copyrights in coaching and working its AI methods.
