OpenAI o3-mini VS DeepSeek-R1: Who wins?
16:41

OpenAI o3-mini VS DeepSeek-R1: Who wins?

Julian Goldie SEO 01.02.2025 25 557 просмотров 533 лайков обн. 18.02.2026

Machine-readable: Markdown · JSON API · Site index

Поделиться Telegram VK Бот
Транскрипт Скачать .md
Анализ с AI
Описание видео
🚀 Get a FREE SEO strategy Session + Discount Now: https://go.juliangoldie.com/strategy-session Want to get more customers, make more profit & save 100s of hours with AI? Join me in the AI Profit Boardroom: https://go.juliangoldie.com/ai-profit-boardroom 🤯 Want more money, traffic and sales from SEO? Join the SEO Elite Circle👇 https://go.juliangoldie.com/register Click below for FREE access to ✅ 50 FREE AI SEO TOOLS 🔥 200+ AI SEO Prompts! 📈 FREE AI SEO COMMUNITY with 2,000 SEOs ! 🚀 Free AI SEO Course 🏆 Plus TODAY's Video NOTES... https://go.juliangoldie.com/chat-gpt-prompts - Join our FREE AI SEO Accelerator here: https://www.facebook.com/groups/aiseomastermind - Need consulting? Book a call with us here: https://link.juliangoldie.com/widget/bookings/seo-gameplanesov12 DeepSeek R1 vs. OpenAI 03 Mini: The Ultimate AI Showdown In this comprehensive comparison, we evaluate the latest logic and reasoning AI models: DeepSeek R1 and OpenAI's 03 Mini (both regular and high versions). Through coding tests, website creation, AI content humanization, and search functionality, we determine which model excels in key areas. DeepSeek R1 outperforms OpenAI models in problem-solving, coding accuracy, and humanizing AI-detectable content. Despite OpenAI 03 Mini's speed, it struggles with accuracy and humanization. Join us as we break down the strengths and weaknesses of each model, offering practical use cases and recommendations. 00:00 Introduction: Comparing AI Models 00:05 Initial Comparisons and Observations 01:29 Testing Speed and Coding Capabilities 04:29 Building a Basic Website 07:05 Humanizing AI-Generated Content 10:33 Search Functionality Test 14:22 Final Coding Test: Snake Game 15:17 Conclusion and Final Thoughts

Оглавление (8 сегментов)

Introduction: Comparing AI Models

O3 mini versus deep seek R1 which is the best and who wins we're going to test it out today first thing we'll do is we'll

Initial Comparisons and Observations

compare the paper on 03 mini and the paper from Deep seek R1 we've actually done a direct comparison inside chat GPT so we can see okay which one is better what's the difference obviously they're both logic and reasoning models an 03 mini and 03 mini High have just been released hours ago so what we can see here is how smart are they and you can see actually if we compare math skills deeps R1 is actually beating 03 mini which is pretty wild and that comes from chat GPT itself for coding they're both pretty much equal general knowledge pretty much equal it seems like writing and creativity they're not really designed for and for Speed it seems like deep seek R1 isn't as fast as open ai3 we'll be comparing them to test them out and you can see there's actually a breakdown here now this is really useful actually the way the chat chbt is presenting it so you can see for hard math and coding help you're probably going to use Steeps I want for fast answers and safe responses 03 mini for deep thinking and problem solving and for a careful rule following assistant 03 mini that sounds exciting so final verdict deep se1 is the best for thinking solving problems and deep reasoning we'll compare them in Practical examples in a minute and 03 a minut is coming out best for fast safe and structured answers and that's comparing both white papers using chat gbt so let's get into the testing now

Testing Speed and Coding Capabilities

and we'll open up deep seek now the model that we need to test is deep seek R1 we're not going to be testing deep seek version 3 because it's all about R1 today and then what we're going to do is we're going to open up o03 mini and we shall also compare 03 mini high right then we can compare all three models all three releases and see which one performs the best so let's get straight into this bad boy we'll keep it simple first of all and then we'll build it up so I'm going to create a Space Invaders game in H we'll ask deep seek R1 first then we're going to ask 03 mini and then we'll ask 03 mini highight now deep seek is still having a cheeky think about this 03 mini has already coded it so way before deep SE go one it's definitely winning on speed and 03 mini High has already come back to us as well so it's a speed test chat CHT an open AI 03 mini is winning by a long way now whilst we're waiting for deep SE G one to actually come up with the response for this post what we're going to do now is we're going to take the HTML and we'll just test it out and see how it's performing so if you copy this you actually don't get the canvas option inside o mini or O3 mini High which means you can't run the HTML inside chat GPT which is a shame but hopefully that comes in the future and we'll clear this out and plug in the response now from what I can see there is no option to actually control the main space invader's character right super buggy so 03 mean High it might come back to us quickly but if it's the wrong answer absolutely abely no point doing that right let's check out the HTML from 03 mini so this is like the mid version we'll test this out see what we got here that's actually giv us a controllable character but again as you can see I'm controlling the enemy it's not the main character and I can't move around so super buggy as well so they might have come back to you quickly but if the answer is totally wrong there's no point right 03 mini and O3 mini High not doing a great job there we still got the HML coming back from Deep C car1 and also just a note here is like you can run this unlimited times inside deep c g one chat gp3 mini and chat gp3 mini high are paid and I think to get unlimited like DC Gan you would have to pay $200 per on the plus flans you get like 150 messages per day so we going to run the HTML now and see what we got here and look at that we're living the dream here Sunshine we're absolutely deep SE kaan has already built something better than space we've SP Invaders versus chat gp03 mini high and chat gp03 mini that has absolutely blown me away because this is free I can use it an unlimited amount I can run the HTML inside the chat I don't have to go to live. com to preview it and it's giving me a better response than the latest version of chat GP which is supposed to be one of the biggest companies in the world and as much as I love openi I love chat GP I love their tools how can it underperform deep C car1 and coding like it just blows me away so Let's test something else out

Building a Basic Website

now um we're going to create a basic website using each model we'll go to new chat we will test some reasoning um we'll create a new chat in each just to make it a different context so we going to plug this in and I'm going to say create a beautiful SEO cost calculator website add emojis plus plenty of details about the tool FAQs Etc brand colors equals this we'll plug that in and we shall plug this directly into deep SE gu one as well and let's see what we get back now again I know that three minis going to go straight Off to the Races and give us a very thorough response in a short amount of time deep seek chat GPT O3 mini high is beginning to give us the output back as well and then deep SE car1 is having a good old think about it I mean honestly for me like if I get the right result I don't mind waiting but let's test these out anyway so we'll go into 03 mini and we have the HTML that's all one single file as you could see so let's copy that and we have the SEO cost calculator back from chat gpt1 mini I mean to be fair it's done a great job there right plenty of information on the page it's nicely designed it's got all the Emojis Etc let's just test putting these keywords in so I'm going to put current ranking 10 website Age Two competition level low number of keywords 10 and that does not work wow that is a really basic task and it's not wored all right let's test out 03 mini High hopefully we get back something better so I'm going to plug that in and we'll test this out now so I'm going to put conversion rate cost per click calculate now and that's worked perfectly to be fair plus we got some nice FAQs so there's definitely a big difference between O3 mini high and O3 mini when it comes to generating websites Let's test out what we've got back from Deep c one so we'll run the HTML we'll test this out we'll just change the number website Pages 15 content Pages five I would say the design is pretty much the same on each of these I wouldn't say any is particularly better for front end design but let's see if the calculator Works which it does beautiful all right so in terms of order and who wins I'm going to go with chat GPT O3 mini High then I would say deep C car1 and O3 mini is right at the bottom right so deep SE caran Saar is outperformed in both tests chat gpt3 mini and it's right up there with 03 mini High however the other thing I would say is just for convenience the fact that you can run the HTML code irly inside the chat here just makes everything 10 times easier so on to the next one let's run new tests

Humanizing AI-Generated Content

now and what we're going to do from here is we just get some AI fluffy content that's obviously AI detectable right now these AI detectors take come with a pinch of salt but I know there's a lot of people who use these AI detectors and it's always something I get asked for right like what's a best model for removing AI detection so if we have a look for example at this blog post it's coming out at 7 77% AI generated right easily detectable so what I'm going to do is I'm going to take this content right here the fluffy content from cha gpc4 and I'm going to test which one or which multiple models are the best for humanizing the content so I'm going to say humaniz is so it bypasses AI detectors 100% of the time must be 100% non AI detectable then I've pasted in the fluffy content underneath that comes out at 77% AI detectable without any sort of humanization so we'll run that through O3 mini High then we'll run it through deepy Gan and then we're going to run it through O3 mini mid and now we've already got the content back I mean the speed of this is just absolutely amazing it is great for Speed chat gpt3 mini so let's plug this content into the AI detector we'll hit detect T see what we get back so just to be very clear here right so oh I've asked for generic AI fluffy content this content as you saw previously came out at 77% AI detectable when we ran it through gpt3 chat gp3 mini it had a good think about it it's like right let's do this it gave us the content back we ran it through zero GPT and it came out at 82% AI detectable that is more detectable by AI than the previous version it's made it more AI detectable not more humanized so let's run Jack GPT 03 mini High through zero GPT is as well we'll see what we get back plug that in and that is coming out at even more AI detectable it's 97% AI bear in mind like o 01 mini which you can't get access to now but 01 mini inside chat GPT would give you 100% non- AI detectable it give you 100% humanized content in my previous tests O3 mini did a bad job let's be honest it made the humanization percentage go from 77% all the way up to 82 03 mini High took the content that was 77% AI detectable and made it 97% AI detectable it did the opposite of what we prompted it to do so hopefully we get something better back from Deep c car one now so we're going to plug this content in we'll hit detect text see what we get back look at that 0% AI detectable wow that is unbelievable that is really unbelievable put personally I W 03 mini to be a great model like I have nothing against chat GPT I feel bad doing these tests because I love chat GPT I love open Ai and what they've done and what they've achieved in life but just to see deep seek R1 which is a free model available for anyone outperforming these multi-billion dollar companies is blowing my mind today it shocks me genuinely that deep seek car1 is just outperforming everything and this is before you know R3 comes Out imagine what R3 is going to be like it's just wild all right next up we're going

Search Functionality Test

to test the search functionality so what we're going to try and do is combine the reasoning model of the AI natively to the search model to see what we get back as well right cuz if we can connect the web and live information to the reasoning model then it can be pretty powerful so what I'm going to say is analyze the market trends in the AI industry based on today's AI news headlines important it must be based on today not other dayss right I want to focus on the most relevant recent information highlight search right now inside mini we'll do the same in O3 mini make sure we select search there and then we'll do the same inside deep seek R1 and we're going to test the search functionality so this is gathering the headlights now o03 Min is doing that as well deep c car 1 seems to be stuck a little bit seems to be thinking a lot there having a deep think so we'll see what it comes back with in a minute and actually deep seek R1 has totally failed right I've noticed this a lot this week like the search functionality is been down so it says the server is busy please try again later one little get around you can do is you can go over to perplexity doai select reasoning with R1 use exactly the same prompt and it will use deep seek reasoning R1 if you're on the pro model you get quite a lot of searches if not unlimited if you're on the free model I think you get like three searches per day and that's working his magic right there so let's have a look what we got back from chat gp03 mini high now bear in mind the headline did say important it must be based on today not other days right and if we have a look at this it's referencing for example January the 29th so that's a bit of a fail although it's done better than deep seek Reuters has so it's referenced January the 29th there to 18th for January 17th of January 2nd of November and also if you read that Outlook it's not really given us like a proper analysis right like honestly if you just watch a YouTube video from today you're going to get something that gives you a better analysis than what you're getting right there on the overall Market Outlook so it says synthesizing these reports the current market Trend in the AI industry is one a transition like you already knew that you didn't need one of the most powerful AI models in the world to tell you that let's have a look at the response from 03 mini so it's still referencing old information even though we asked it not to so January the 18th January 17th it's talking about really generic stuff and I would afford that one of these models would find like the latest information from Deep seek all the craziness that's been happening this week you know referencing the information from the last couple of days Etc especially if you know it's focusing on today's news headlines like for example if we didn't use Ai and we just use for example like Google which does leverage AI but if we went to the news headlines we can get today's headlines just by doing that right and that gives us a much better output so it's pretty disappointing in terms of the results from 03 mini high and also DC gu one let's have a look what we got back from perplexity maybe it's something better so look at that perplexity is probably outperforming everything especially when you combine it with deep SE car one it says the AI industry is undergo undergoing significant shifts as of and then it actually mentions today's dat which is useful then it talks about deep seek which has been the biggest headline from this week the thing that you would focus on if you were doing an analysis from today along with that is talking about how Taiwan flag deep Seekers a national security threat along with Italy Banning its AI platform from App Stores this is really recent news right and the analysis is super useful so out of everything perplex he winning that especially when you combine that with deep sear one game changer let's do a final test so I'm going to say now

Final Coding Test: Snake Game

bill being a snake game in HTML as always O3 mini is ready fastest so let's copy this we'll test out and that seems to be super buggy look at that it's like flickering all over the place so for just coding in general it seems like O3 mini is not that great let's test O3 mini high again look at that super buggy like it's just flashing around when I try and move it not great at all deep C car1 is still having a good think about it a good deeper thing that's what we like to see actually I don't want a Recon model that comes back to me too quickly cuz it doesn't seem like it's fought it out very well you can't rush greatness especially a snake game so we're going to hit run HTML like so hit start game and look at that deep seek never lets you down there we go look at that so as you can see the snap game the only AI model that was capable of building the snake game with very simplistic prompts is deep seek R1

Conclusion and Final Thoughts

and so out of all the models I'm still going to stick with deep seek R1 if I need logic if I need reasoning Etc I would stick with that model if you're worried about privacy and all that sort of stuff I would just install it locally run it on an offline laptop o mini high and O mini not they've not blown me away at all they've let me down a couple of times today in terms of humanization building out games building out websites It Works simply enough and for searching the web and research and using the power of reasoning with research I would just go with perplexity plus deep C1 that is free and as you saw from today the results were really good so thanks so much for watching if you want to get access to a community of 3,700 people interested in Ai and feel free to get that link in the comments description you'll get a free Aro course plus 5 fre Aro Tools Plus 200 free chat GPT Aro prompts links in the comments and description and if you want to get a free one to1 SEO strategy session that shows you how to get more lease traffic and sales to your website to grow your business to get more customers feel free to get that on this link building acceleration session you'll get a free SEO domination plan so you'll get a custom tailored link building plan so you can generate more lead sales and profits from your website you'll discover the secrets of SEO link building will answer any questions you have want one you learn the best Ling bu strategy for your website plus add to quickly our ranking compe Ling now to dxo traffic based on what's working for us and our happy clients feel free to get that link in the comments description appreciate you watching as always byebye

Другие видео автора — Julian Goldie SEO

Ctrl+V

Экстракт Знаний в Telegram

Экстракты и дистилляты из лучших YouTube-каналов — сразу после публикации.

Подписаться

Дайджест Экстрактов

Лучшие методички за неделю — каждый понедельник