Deep Research Agents: Google VS Perplexity VS Grok VS ChatGPT...Who Wins?

so Google versus perplexity versus Gro versus chat gbt deep research agents

Google just released the Deep research update for free across the world right so you can get free access to that super powerful stuff let me show you an example of what the report looks like so you can see an example right here so here's an example report right 5 half th000 word report on the difference between these different agents one of the most powerful things about this is like it integrates with Google Drive so you can export it to Google Docs and also this completely free you can do it in many different languages and I would say it's made Gemini a whole lot more powerful also this is using the latest model of Google which is Google Gemini flash experimental 2. 0 so that's what it's using to generate these reports and that's why they're so good so we're

going to be comparing the outputs versus Gro chat GPT perplexity Etc one of the things to note here is that there are different limit on each platform right so for example Gemini if you're using deep research is free grock is free up to a limit chat GPT previously it was only available for Pro for operator so for 01 Pro mode users plexity is completely free as well and of course they all use different ways of generating responses as well right so for example perplexity you can actually change your settings here if you go to settings then if you scroll down you can use different AI models to generate so for example like Claude or that sort of thing if you're on the pro version of course using Gro and chat CH so let's compare I've got a bunch of pre-made

examples right here so this is the report from Gemini like you can see 5 and a half th000 word response and also what you can see here is it's used a ton of different sources so these are all the sources used within the report and additionally this is pulling in like the latest information from what I can see so for example this report here was released on March the 13th and it's pulling in the latest information now I've done exactly the same prompts inside chat GPT and here's an example right this is the response we got back from chat GPT so we can compare them side by side let's check the word c first of all how long is this word now this is long report so we'll check the word count on this and there was a couple of big errors I noticed inside report from chat GPT so I'll come on to that in a second but yeah this is what 19,000 words chat chpt created a 19,000 word report let me double check the word count of this report from Google so that came back to us with a 5,000-word report I also created the same prompt inside perplexity using deep research so if you want to use a deep research inside thep Plexi just go to new thread and then select deep research like you can see right here and this is the Deep research report back from perplexity now one thing I noticed from blex is it tends to be a lot less bur than the other platforms so this is only 1,200 words on the research report but in a way I prefer that am I going to spend a day reading through walls of text like this probably not right do I need a 19,000 word report comparing all the differences between them probably not and I'll show you why this actually has a big problem inside the report I'll come on to that in a second but yeah I think that perplexity actually is got a really good balance between being fast and actually getting the outputs that we want and then finally we have the report from Gro as well now gr's reports are a lot nicer formatted right so we got like nice bullet points here quick overview get straight to the point it's not messing around and then we have a breakdown of functionality Etc so this is much easier to read are nicer formatted than any of the other reports so if you compare this report versus Geminis which one is easier to read for sure it's Gro and then in terms of usability I'd say Plexi comes in second on that but this one is probably the best right so far however if we go over

to chat gbt so we use the same prompt in every single one right which was research the differences between Google versus perplexity versus Gro versus chat gbt deep research agents right and the key word here is deep research agents now if we scroll down and we look at the outputs this report number one it took the longest to generate out of all of them so out of all these reports chat GPT actually took the longest to generate and number two if you actually look the outputs here you go through it with a fine tooth comp you can see here that Google is referenc as having Google Search right so it's not actually focusing on deep research agents which it should be it's just focusing on Google Search and the same for example with perplexity so focuses on perplexity AI search not deep research agents so this might have given us like 19,000 words of reporting but what's the point if it's focusing on the totally wrong model right it should be focusing on deep research agents now if you go back to Gemini for example and we have a look at the differences here we'll compare them inside here so you can see that Gemini was smart enough to figure out that we're talking about Google deep research and the same for Plexi deep research right that's a key word here so chat gpt's comparison is pretty much completely wrong it doesn't matter how long it is how many resources he used it's pretty much hallucinated because it's focusing on Google searches not deep research and so for me personally out of all these models I'm going to say the chat GPT actually comes in last right now if we go back to Gro we'll see what we got here so we've got the examples and again he focusing on Google AI search I want to make sure that it's actually FOC yeah so you can see here it's actually researching using Google AI search not using deep research right so gr has failed that test as well because instead of being smart enough to figure out okay this is about the Deep research model inside Google and inside perplexity instead it's focusing on the general search methods right so the research here is pretty much wrong as well and so for me personally I'm going to put chat chpt and gr three deep research models has last right they've completely misunderstood the search intent that we're going for and because of that the research it's given us back is totally wrong not very good there all right now we'll keep going and we'll compare the rest of the research models here so we've got the report back from

perplexity it's comparing the four models it understands that we're talking about deep research models and you can see here it specifically references Google describes it deep research tool as an agentic feature right so it understands that number one we're talking about deep research not AI search or Google search in general and number two it's sourcing a report from very recently right from January 2025 so it's smart enough to figure out the searching turn and okay here's what you know we're looking for and here's how the report should be formatted so he's definitely coming in second or first now let's have a look at the comparative analysis so one of the things that you don't get inside this report using the Deep research model is you don't get a table breakdown right so if we go over to Gemini it's giv us like a really powerful deep research breakdown of all the different models right and we can actually export that to Google Sheets which is super useful as well whereas perplexity search it doesn't do that right and so if I was to compare the AI agents for deep research versus perplexity versus Gemini just having that table makes everything 10 times easier to understand so I'm just going to format this better right here and then we can compare them let's pull this up all right so how much easier does that make it to understand the research you got a nice little report here it tells you the pricing differences in accessibility research methodology Etc you just don't get that from complexity and comp in each of these for me Gemini is winning and bear in mind as well Gemini is completely free Plexi is completely free gr is limited by premium limits and chat GPT is limited as well plus chat GPT takes a long time like for example if we go to a new chat here and we'll ask it to do the same thing again so let's pull up a new example we'll select deep research then we're going to plug this in hit enter number one is going to clarify what you're looking for and then you can say yes go ahead and then once it comes back to you can see how long this task is going to be right so that's probably going to take about 20 to 30 minutes to get back to us whereas Gemini was a lot faster I think Gemini generated in about 2 to 3 minutes so the differen is in speed as well just absolute night and day between these different AI agents now if we pull these up side by side let's compare the differences here so it's got a breakdown of the underlying model so for deep research this is using Gemini 2. 0 flash thinking experimental flexi is using it AI search engine Grog deep search is using Gro 3 and chat gbt you can switch between the models if you want to now this is not 100% perfect because actually when you're using the Deep research model on chat gbt it says specialized O3 model but actually you can use any of the ones you want to use right so if we click on new chat over here we can switch between these and then we can use deep research on o1 or GPT 4 out so the report is still not as good as a human right it doesn't have the same level of expertise as someone like me I'm not trying to do a humble brag there but you get the point and then with Google deep research as well pricing is free with chat gbd deep research it's on the $200 per month plan and these to be fair this research is not quite right as well so onlex is actually free and on Gro deep research is free as well chat cheap I'm not sure if they've extended it for free but yeah so there's definitely a lot of mistakes in all these reports it's any however you generate the Deep research you're going to want to double check it before you send this especially if you're doing it for like an academic situation or you've got a report to create for people that you're working with you definitely want to check this stuff before you send it off otherwise you're going to give people like the wrong information and it's going to backfire it's going to do more harm than good right but yeah I would definitely

say out of all of them Gemini is winning there followed by plexity these two are free chat GPT is on a Premium plan and it takes a lot longer to generate so you can see here it's still generating reports here and then Gro as well is pretty much wrong so yeah the other biggest problem with these agents on chat GPT and perp Plexi sorry chat GPT and Gro is that they don't understand she's talking about deep research right so it's not really understand the query I inserted that cuz I wanted to see how recent it could go back with the data I also want to see the sources uh This research has used so if we go back to the output here let's see how recent some of these sources are so we'll open up some of these see what we got so yeah it's referencing like the main model of Google Gemini is it referencing the latest blogs from them yeah so it's two or three days ago which is awesome as well yeah I think if you're going to do deep research just go with Gemini that's the best set of all of them so thanks so

much for watching appreciate you watching if you want to get access to all the notes and the comparisons from today feel free to get that inside the sap section of the AI profit boardroom I've got a section on Google free deep research agents here I've also got a three-hour course on how to build anything with maners so this is a three-hour course with all my best SS if you can't get access to Manus then you probably want to check out the Alternatives I've set up with recipes right here and this also comes with all my best workflows tips AI agents Etc and you also get all my best training on AI automation like you can see if you want to scale your business make money and save hundreds of hours this is one of the best communities you can join and prices are going up at the end of the month Additionally you can jump on weekly Q& A with us so the weekly Q& A calls with us here and if you have any questions you can post inside the community ask our community and we have 465 people ready to help you like you can see right here so for example Chris he posted a couple of days ago he's already got the answers three people have helped him out there and so if you're looking for help like this is one of the best places to ask and you can just get answers whenever you need them so feel free to get that link in the comments description to the AI profit boardroom additionally if you

want to get a free SEO strategy session feel free to get that link in the comments description and with this free SEO strategy session we'll show you how we de websites from 0 to 145,000 visits per month and generate hundreds of thousand of do in sales on Al pilot on this free link building acceleration session you'll get a free domination plan so you get a custom tailored link building plan so you can generate more lead sales and profits from my website you will discover the Secreto link building or answer any questions you have you learn the best link building strategy for your website plus how to add ranking competitors and 10x SEO traffic based on what's working for us and our happy clients like you can see right here feel free to get that link in the comments and description appreciate you watching and I will see you on the next one