quen 2. 5 Max versus deep seek which one wins today we're going to be testing out both AI models because there's a brand new release from quen which is quen 2. 5 we can get free access to it today and I'll be running side by side deep SE versus quen 2.
5 to see which one performs the best now if you want to compare the benchmarks right here we've got them lined up uh you can see here that quen 2. 5 Max is actually beating deep seek version 3 in many of the benchmarks now this does not compare deep seek R1 because R1 is more of a reasoning model but we'll be testing out deep seek version 3 versus quen 2. 5 Max and seeing which one comes out on top today it's going to be pretty interesting stuff now one of the best things about quen 2.
5 Max as you can see right here so you just go to chat. lm. and then select Max in the top left here is that you can actually switch between different modes so it can create images it can code it can make plants it can generate videos and even Imes which deep seek currently cannot so for example we're going to unclick deep seek G1 cuz we're not testing that model today we're testing deep seek version 3 maybe we'll look at them later on but basically we're using deep seek version 3 versus quen 2.
5 max they're the most comparable models and we'll see which one performs the best so let's kick it off and see what we got here first thing we're going to do is we're going to test out content writers so I'm going to say create an SE optimize article for this and we'll put deep seek versus quen 2. 5 Max we'll delete that we've given it a bunch of instructions for creating the content and additionally I'm just going to insert some relevant personalized information about us now we'll be comparing the exact same prompt inside deep seek version 3 versus quen 2. 5 Max so let's hit go on both of these here we go and what we can do is pull these up side by side so let me do that for you so quen 2.
5 Max is over here and deep seek version 3 is over here now if I had to compare them side by side I'm looking at the titles deep seek versus Quin 2. 5 Max they've both front loaded the keyword however I would say the actual writing is far better on quen 2. 5 Max and look at this right so if we look at this teaser here it says let's break it down in a way that's easy to understand conversational straight to the point it feels weird like why would you even explain that you just get straight to it whereas for example if we look at someone who's been in the trenches of SEO and AI for years I know how overwhelming it can feel to pick the right tool let's break it down together no fluff no jugging just straight talk both of those are probably not as good as claw 3.
5 Sonic when it comes to writing but I would say 2. 5 Max has the edge now the other thing I like about this is inside Quin it kind of has this canvas feature where you can select the content and then you can ask it or explain it to do different things whereas inside the editor for deep seek you can't do that you can't edit it like a canvas operation right so for example if we select that there's no like ask or edit option so if we click on this then we click ask and we just say make it shorter this kind of chat gbt canvases feature where it's Rewritten that section of the content and we've got a better version right there so it's easier to edit the content inside quen 2. 5 Max as well very impressive honestly for test number one content creation quen 2.
5 Max is is outperforming deep seek version 3 by a long way right there even if you look at the first line right let's cut to the chase versus hey is Julian Goldie here no one cares if Julian Goldie is here or not if you're going to write an article you want something to get straight into it right so now we're going to test see artifacts teacher inside both so we'll start a new chat for both of them and I'm going to select artifacts right here and we're going to say create a space invader's game and we'll do the same inside quen 2. 5 Max and we'll test the coding capabilities now if you using deep seat normally you probably wouldn't be using deep seek version 3 for a coding test like this instead you would select deep seek R1 but to make it a fair test and compare the most relevant models together we're going to stick with deep seek version 3 and quen 2. 5 Max now the first thing you're going to know is that this is writen in Python whereas this is written in HTML what does that mean what that means essentially is it's going to be much easier to preview the game inside quen 2.
5 Max than it is inside deep seek so for example you can't run or preview python code inside the chat of deep seek version 3 right and so how do we know if it's good what we'd have to click away we'd have to go off we'd have to test it run it on Terminal Etc lot of messing around there if you're not someone who's good at coding like me I don't have any coding experience if you go over to quen 2. 5 Max we've got the preview ready to go on the right hand side over here we can click on artifacts that's going to load the game and play the game just as you can see right here in literally one prompt right pretty mindblowing that you can do that and so Quin 2. 5 Max on coding as well is winning too so all we literally did was say create a space inv this game it created it we can preview it in artifacts and it actually works boom sha lacker if you go back to deep seek it's created in Python so I can't even check if it works inside the chat so just for ease of use and convenience this might work beautifully but for ease of use I'm going to say coin 2.
5 Max is winning so two out of two four Quin 2. 5 Max of up and by the way if you want access to deep seek you just go to chat. deep seek.
com all right so let's try another thing now we're going to say create a beautiful SEO cost calculator website add emojis plenty of detail about how it works Etc and I'm going to hit enter on that and then inside 2. 5 Max we're going to use the same prompt hit enter and start coding that now I will say if you compare them side by side you can see here the Deep seek tends to code faster but quen 2. 5 Max is not messing around it just gets straight to the code whereas inside for example deep seek it's responded faster but it's not actually coded at all so far let's see if it comes up with HTML so it's explaining the website is create the content create the copy Etc but so far we've not had any code back all right and it actually hasn't created the the website it's just like written the text written the content but there's nothing in there that actually has a HTML whereas inside quen 2.
5 Max we have the HTML ready to go you know it's not messing around on the response there it's not giving us like a fluffy BS answer it's just like let's code your tool out for you boom shakalaka and it did the same with the Space Invaders option as well because it's specifically Target in AR because it's specifically got a option for artifacts it's like this man right here who's prompting me he just wants HTML he's not messing around today and so you can see how it just goes straight into that so for coding apps like Space Invaders or games Quin 2. 5 Max is winning for creating content Quin 2. 5 Max is winning for creating websites quen 2.
5 Max is winning my friends and you can see that's a pretty nice option right there so we can select between packages let's test this actually works we'll put 100 keywords in there boom works perfectly we've got a nice little explainer on how the tool works frequently asked questions that's exactly what we wanted and it did it in a one take Drke now let's test them out for creating content and right so what we're going to do here is I'm going to say to chat GPT generate some generic AI fluffy content all right and basically what that's going to do is just create like a super generic piece of content that we can then try and humanize so for example we'll just take this content we'll copy it we'll paste it into Z GPT just to make sure it is AI detectable that's coming out of 100% AI detectable so now what we want to test is can deep seek or quen 2. 5 Max humanize his content properly we'll start a new chat we'll paste this in and the prompt I'm going to use for this by the way all the prompts are inside the fre SEO of course Link in the comments description this humaniz so bypasses AI detectors 100% of the time must be 100% non AI detectable now if you use something like deep C car1 or chat gp01 pretty much every time you get 0% AI detectable inside zero GPT with exactly the same prompt but let's compare deep seek version 3 versus quen 2. 5 Max now let's test this out I've taken the content from Deep seek version 3 plugged it into zero GPT and you can see it comes out at 94% so that score is only dropped by 6% directly from Deep SE version 3 which is not great it's not ideal let's be honest so now what we're going to do is we're going to take the content from quen 2.
5 Max we'll plug that in and see what it comes back at and that is even worse right so quen 2. 5 Max has underperformed formed versus deep seek version 3 for humanizing Content now here's a little Pro tip for you if you want to humanize content select deep seek R1 use exactly the same prompt and use that to create the content all right so if we take this content right here that's coming back from Deep seek R1 like you can see we'll copy that we'll go to Z GPT just like magic my friends just like magic deep SE car wand 0% detectable for the fact that deeps actually has our one built in it's completely free I would say it it wins that test easily so let's start a new chat now so what deep seek doesn't have that quen 2. 5 Max does have is image generation so let's say for example we'll say create a picture of a cat eating bisque all right and then we head into like and that is pretty fast for generating the image you can see that's just absolutely flying through the image generation right there obviously you can't compare this versus deep seek Because deep seek has no image generation but that is looking good I did look at Janis Pro 7B which is the Deep seek model for image generation and ran the same prompt yesterday and check out my video on YouTube it didn't look anywhere near as good as that that is impressive to be fair that's very good let's try out video generation now so we're going to start a new chat again we'll go to video generation we'll put cat eating biscos and then we'll head send the message and we'll see what's coming back so this seems to be a lot slower I don't seem to have a response yet let's see what's going to happen next now seems to be responding obviously it's going to take a lot more power to generate videos right so you can see how slow this is whilst I'm talking we might have to come back to that in a second whilst we're doing that let me open up a new tab of Quin 2.
5 Max we'll start a new chat here and what we're going to do now is test the search version of both right so both of these can do web search uh what I'm going to say now is what are today's AI headlines and we'll plug that directly into qu 2. 5 Max along with deep seek and we'll see what we get back back now the biscos video is still generating so we'll come back to that in a second and look at that deep seek search service is busy please disable search we'll try again later right so we can't use the search feature inside deep seek quen 2. 5 Max has come back to us and it's given us the headlines for the 28th of January which is about 24 hours ago but still doing pretty good there we're on the 30th of January this gave us headlines for January the 28th it's okay it's not today's headlines like we asked it for you can see a lot of these are quite outdated so for example we got like 2024 December 17th but it gave us a response which we didn't get back from Deep seek as you can see it struggled too busy it's been a crazy week so I understand but still which one won the test it's going to be Quin 2.
5 Max so for image Generation video generation web search artifacts content creation Quin 2.
