ChatGPT Agent Mode vs. Genspark Super Agent (Here's The Winner)

Speaker A: I use ChatGPT agent mode and Genspark super agent side by side for everything and the winners shocked me Today I'm going to walk you through three simple examples to show which agent is superior. For those of you don't know me, my name is Ryan Staley. I have helped thousands of sales and marketing executives with AI transformation and I'm really big die hard when it comes to different AI tools. What I'm going to show you today is a side by side comparison. So super hyped you when agent Moat came out. And then at the same time I've been using genspark more and more. So I wanted to really test them side by side to show you what's real and what's happening now, what's available. Like I said, you're going to be blown away. But most people are not super familiar with genspark so I want to start off with that and give you a super quick background on who they are and then I'm going to get right into it and share. This is going to be a quick video like normal, but I want to walk you through through who they are and what they've done so then you have some context as it comes to it. So one of the cool things that you're going to see here is um, and this is a piece on off of lack a database but they went from 0 to 10 million faster than OpenAI. And this is really, really interesting. They did a sprint and I mean in nine days went to 10 million 10,000 paid subscribers. The monthly price is $99. It's actually only $20 now and their current valuation is 530 million. One of the things that's amazing about these folks is, you know, Eric, who is the CEO and founder, seems like a really good person from when I've interacted with them. But one of the things that's really interesting is they're now close to 36 million in revenue, some even say 45 million in revenue. And this is only 90 days after they started. So I'm going to have him on the show, him or his co founder on the show and talk through through some more specific details. But what I wanted to do is really infuse genspark with what's going on there with really what's happened with OpenAI agent mode. So I'm going to switch over to here and what I did is I gave these models side by side tests with the same exact prompt with no change so you could truly understand how they're performing, what the output was the quality and the price. Okay, so first things first. Uh, basically ChatGPT agent mode, which is new. Think of that as like a browser based agent. They could click on different pages and examples to come up with a creative outcome. Right. So it'll go from like searching the Internet. Think of it as like a more advanced version of operator mode. And so in this situation, I asked them to create a go to market strategy like McKinsey for B2B SaaS company at 100 million in AR that wants to 2x in 3 years have this specifically designed for AI agents for B2B sales and marketing teams ranging from 50 million to a billion. Um, include visuals, Personas, pricing, roadmap and market insights. All right, so this initial test is more on. Okay, what's the output it can create from a presentation perspective. And what you're going to see here is it's really simple to use. Um, all you have to do is be an O3 and then click on this agent mode. As you can see, I have 386 left. Um, you do have to have the $200 a month user license for this, the pro user. So that's um, something to note on. But, but let me walk you through what I came up with. And so this is the presentation and I'm going to make this bigger so you could see this in full. Okay, so this is the format and this is the presentation it created. And here's the interesting. It took about 30 minutes to create this. Okay. Um, when I used Genspark it literally took about four minutes. So that was something that was really eye opening to me. And then here's the slides. As you can see, they're okay. Um, I'm not really blown away with these visuals. This kind of reminds me of like when you ask PowerPoint to just create the slides. Um, and there's some really, I mean there's some semi interesting stuff here. But like I said, this is very, very high level. So I wasn't blown away with this output. And what I'm going to do is I'm going to show you the output from genspark. And as you could see, I mean even if we do this, I mean you're going to see the big difference immediately. So if we look at genspark, what you're seeing here is like this is really hyper detailed. Let me go, actually let me go to version one so you could see this because I had a additional, um, version that I said, okay, make this 10x better just so I could see what it looks like, um, I think this is okay. As you can see, it said here, see the presentation. Okay, so actually let me walk you through this. This is probably better to do so with this. What you're going to see here is basically it created an outline of a plan before it did anything across, like I said, same exact prompt across 15 slides and then it created those slides. It's using all these different tools and shows you the example of what to do with this or how it's doing it, I should say. And then it covers as a summary the key sections, highlights all those different areas. So that was something that was really interesting. And then like as we go through the visuals on this, you know, one of the things is that it lists the table of contents. Like, uh, look at this compared to like the previous version, right? I mean, of what OpenAI did. I mean these visuals are really, really, really impressive and the data looks a lot more solid. But I was really, really impressed with the outputs that this model created. Okay, Like I said, this is pretty close. For some reason it's not pulling up. The, the initial version, it was pretty close to this. I said make it 10x better and it added a little bit more color, a little bit more differential in terms of the presentation. So for test number one, I would definitely say Genspark killed OpenAI agent mode. In terms of the presentation. It was really operating at a $20 license. This is what I did. Um, it took probably about four minutes max. OpenAI, uh, took about 30 minutes. And then in the next step, basically what it did is, um, it also created a much better quality. Okay, so now what we're going to do is we are going to shift back to OpenAI and I'm going to show you test two. So test two, remember the thing that I did was I'm like, all right, how can we use their agent? Right. So, um, more the general purpose agent. And so I gave it this prompt specifically to design. And this is a mega prompt, right? So this is around creating a YouTube video. So I'm going to share this all the way at the beginning. But as you can see, this is kind of like a complex task. So having it act as an expert level research analyst, narrative strategist and audience engagement specialist. So as you can see, I have phase one here, which is foundational research. Um, as you can see, phase two is world class YouTube video outline creation. And phase three is more around kind of the details behind it. Right. The purpose. Uh, where's, where's phase three? Oh, it's got to be on here somewhere. Don't. Don't skewer me in the con. Okay. Yeah, this part's long. I forgot. I'm sorry. Um, okay, phase three, Discoverability and polish elements. This is everywhere with keywords, narrative, video, description points, all those areas. Okay, Once again, I did the same exact prompt in both of these. This was the output. Now what happened was it worked for five minutes and then it stopped and asked me a question. Um, at the same time I said continue. It stopped after seven seconds and then I answered the question. It thought for 38 seconds and then it finished it off. So what you're going to see here is it's got foundational strategy elements. Uh, it's not just the next model. Because I did this on GPT5. I'm like, all right, let's do it on GPT5. This is probably going to be one of my upcoming videos. Now there's some pretty good content, highlights, visual, visual and audio cues. Um, really interesting. The smart thing I liked about it is it, it starts telling me to point to other playlists. I have my channel, which is really sharp. So, uh, at the same time they gave title ideas. ChatGPT, uh, explained one model rule them all. Some pretty good, like YouTube titles, I think, in my opinion. Uh, and then it has the key bullets, keywords and tags, sound and then key sources mentioned on the screen. All right, now the one thing I was like, all right, how can I test this agent one step further? I'm like, all right, now create a deck of visuals I can use for this video. Now I left it in the agent mode. So I did not Change this to 4.0. Basically what it did. And I was a little disappointed with this. It just created like a description of it. It really didn't. I mean it just, it gave it like kind of a play by play, like, here's the strategy. But it didn't really do anything. Okay. And it said, your visual deck is ready in the canvas. But it's not a visual deck, it's just an outline for a deck. All right, so now what we're going to do is we're going to pop over to genspark and see how they perform with genspark. Like I said, I gave literally the same exact, um, prompt. Now, oops, this is the wrong one. Oh, Here we go. Chesspark 5 release date. So what you're going to see here is. This is what I went through. Literally the same exact prompt, super long and detailed. And now what you're going to see here is, um, it Went through using multiple tools again. And what I will tell you too, in terms of time, uh, I think OpenAI's, like I said was like five minutes worth of thoughts. So then finished off another, so close to six minutes. Uh, genspark did this in, I think it was like less than two minutes. Right. So one third the time. Uh, at the same time, I'm looking at the quality. There's some pretty interesting things in here, like shocking leak million token context window rumors. So there's some pretty cool things that are within here, uh, in terms of the research and it's not as long as what OpenAI had, but there's some really, really important things that I think will jump out. Right. Um, it had some interesting hook elements that it really focused on, which is one of the most important things. It's like 80% of attention with YouTube. Uh, and then it goes more into the whole entire script. Okay. So in 10 minutes you'll understand why GPT is just OpenAI's next model. It's the moment AI stops, become a tool, becomes a competitive advantage. By the end of the day, you'll know how this affects your business strategy in the next 90 days. Or cool. It's got the context. Um, I'm going to go through this different segments, right? Uh, which is, uh, I mean, this is really good. And now what I will also say is before going through all these semantics, it means even got end streams and end screen strategy, right? It's got discoverability and polished. So it's got compelling title options, video options, keywords, even sound in source distribution, and then final execution notes. Now I asked it the same question. I'm like, all right, let's see if this will go multimodal. I didn't switch modes. I didn't go to a different agent. I said, can you now create a deck of visuals I could use for this video? This is where, once again, I was impressed with genspark. And it just asked me to confirm, like, we got hook, timeline, process, roi. And not all of these are perfect. However, there's some usable things in here. Like, I like this one. I thought this is pretty cool. Like, this is like a confidential, like, visual that I can include within the file. Like, and it even had today's date on it, which was cool. When I recorded this or earlier date, this thumbnail is actually pretty good, I thought, right where it's got like all these different models and it's hitting the multimodal with GPT5. I thought that was like, pretty clever. Now some of these this one didn't operate so good because the text is all blurred and same here, it's uh, more gobbledygook. But what you can see here is they continue to create, try to create all these different visuals, create some more thumbnails. They use multiple different model tools. Funny enough, the image tools they use were the ones that were GPT4O and GPT didn't even use that 4.0 image model. So that was pretty interesting to me. And then it basically says, okay, here's where I thought it was pretty clever. It says like visuals to use for different sections, uh, as it goes through. And now like I said, all of these aren't good because like this one, it's context window size, like this is sputnik, like it's just like, just not even real words. Um, so obviously I can't use that. Um, but I think the first few parts of them were pretty good and then it finished with some like next slides. So um, what I would say is for round two as well, Genspark one man and one on like the Super Agent, Genspark, Super Agent versus Agent mode and OpenAI. So now that we're going to go take it one step further, I'm going to take you back to Agent mode and chatgpt and we're going to show you what happened when I asked it to create a one pager. Okay. And this was really simple, uh, on um, what I asked it to do, I said generate a visually stimulating one pager of me and my offering based on my LinkedIn profile. Uh, exude professionalism, leadership and AI forward thinking. And uh, then based on, I gave it a brief description, then I gave it my link. So very simple, nothing crazy or over the top. And uh, I said okay, let's use the LinkedIn profile. It worked for 8 seconds and then it's like asking me more questions. Um, it's kind of interesting because genspark doesn't do that. Um, and I've noticed when I work with Google Deep Research as well in Gemini, it doesn't do that either. It tells you the plan. Uh, so this is something that OpenAI clearly has built in and basically what it does is it created a personalized one pager. This is the one pager that it created. As you can see the formatting is off my name's off the screen. Track records and achievements. Not a very good look uh, for ChatGPT when it comes to this. Now if we put Genspark back on, I want to show you what happened with them. I uh, really basically had the same exact prompt. Not basically, it's literally the same exact prompt up here generated visually stimulating one pager. As you see, it's exactly the same. And this is what it came up with, came up with that, um, has an executive summary. It has some different things that I've done in the past. What I would say is this isn't perfect but for the limited prompt that I gave it, it was pretty interesting. Now here's where it blew me away. It said what leaders are saying. So these were basically posts that people had. I did an AI augmented GTM course where I did this for 900 people. So there's a lot of go to market leaders like sales and marketing executives are talking about this on this. Basically it found those quotes on the Internet in social and then put that on this one picture. Okay, that's super creative. Um, it also talks about my core expertise in terms of 3 to 5 Xing productivity architect, ah, go to market systems and reimagine sales. And then it even talks about a five step methodology for AI augmented transformation. Um, so even some results here and then credentials and affiliations. How m. I'm part of the OpenAI community. And then same, uh, pavilion course creator for the AI augmented team talks about awards. So like I think this blew out uh, of the water what OpenAI did with Agent Mode. And here's the thing, the weird thing about this is I'm literally one of the biggest fans of OpenAI in terms of what they're doing. Sam Altman and the team, I think they're typically further ahead of everybody else by three to six months with what they're doing. However, with what I'm seeing with OpenAI, agent mode versus Genspark. I mean Genspark absolutely crushed it with this and they're a fraction of the size. So one of the things that uh, I want to share with you all is like this is what I think is going to continue to happen, not just with OpenAI but anybody else. You try and cover too many things in too many different areas fast. You have the potential to run into qualitative challenges when it comes to this. And so as you can see, I know this is still in research preview for Agent Mode. It's got massive potential with OpenAI. But uh, all the times that I've used it, it has not really gone that well. It's been very, very mediocre. Now it's got potential, right? But really impressed with what genspark did here. I thought they crushed it. I like how fast. They're shipping and putting new products and new features out and they're absolutely obsessed with this and created an agent for Vibe working and so wanted, um, to share this video with you. As you can see from my new format, I kind of alternating or having either an expert showcase their AI tool on here, or I'm walking you through something specific. So I'm going to continue with that format. I'm trying to get the, like I said, the CEO or one of the co founders of genspark Onyx, I think they would be great to talk through and show this and do a deeper dive. But I wanted to share this with you because one of the things that I think is a big disservice that I'm seeing on social media right now is there's a lot of people and influencers that are throwing things out there saying this is the best since sliced bread. When I actually try it out, my experience is completely different. I want to share that with you so you don't need to spend hours and hours and hours doing that yourself. So then eventually, through using it, following my videos, implementing what I'm talking about, you could 10x, even 100x what's possible. So thanks for joining me today and we'll see you all in the next video.