Openai o1 vs Deepseek r1 | Who is the best Coding Model?

Поделиться
HTML-код
  • Опубликовано: 29 янв 2025

Комментарии • 80

  • @rexmanigsaca398
    @rexmanigsaca398 3 дня назад +44

    DeepSeek is the true "Open AI". It will crash NVIDIA stocks soon. With DeepSeek, it shows that you don't really need all that compute power to achieve the results that you want.

  • @elvisadrianable
    @elvisadrianable 3 дня назад +23

    The dragons of GPT are just pure comedy .... it even adds text "behold the dragon" and it looks like a potato looool. It's been a while since I laughed so hard

  • @dubesor
    @dubesor 3 дня назад +23

    I really like DeepSeek. Such good models. I completely switched for now, V3 for fast interactions and R1 for longer more complicated tasks, it's a powerful toolset. Also my API cost has gone down by 80%.

  • @u007james
    @u007james 3 дня назад +29

    openai o1 seems abit lazy

    • @user-ze9tj9yj4t
      @user-ze9tj9yj4t 3 дня назад +3

      yep, you have to write really detailed prompts with complicated wording to get a good result

    • @u007james
      @u007james 3 дня назад +1

      @user-ze9tj9yj4t sounds like how bard started

  • @infofleare
    @infofleare 3 дня назад +9

    Happy Chinese New Year. It is free for everyone.

  • @FreqtradeFR
    @FreqtradeFR 3 дня назад +6

    Amazing comparison between DeepSeek and OpenAI - that animated dragon from DeepSeek was incredible, and it's awesome to see how both models handled Mario and Pacman! 🎮

  • @ehza
    @ehza 3 дня назад +1

    nice work dude...keep it up

    • @YJxAI
      @YJxAI  3 дня назад

      thanks man

  • @pixelpete94
    @pixelpete94 День назад +1

    How are you running the code what software are you using for this?

    • @YJxAI
      @YJxAI  День назад +1

      Both the websites can run html and react code natively. Chatgpt has canvas and deepseek runs it when it identifies any code similar to that.

  • @memeherp166
    @memeherp166 3 дня назад +16

    then why should i pay gpt ?

    • @rexmanigsaca398
      @rexmanigsaca398 3 дня назад +6

      Already unsubscribed from ChatGPT

    • @roro-v3z
      @roro-v3z 3 дня назад +4

      I unsubscribed 😂😂

    • @soumen08
      @soumen08 3 дня назад +3

      One reason could be that for useful tasks, o1 is better, and another reason could be that you do not have to hand over your data.

    • @mckane6989
      @mckane6989 3 дня назад +3

      @@soumen08Can you elaborate more on the handing out data part?

    • @10minuteimprovementpodcast
      @10minuteimprovementpodcast 3 дня назад +2

      practically the only thing they have that others don't have is that advanced voice mode which is speech to speech AI but even that is limited to an hour a day which is a very cheap move

  • @SakshamBhattarai-z9k
    @SakshamBhattarai-z9k День назад +4

    6:28 bro gonna start ai war

  • @davout5775
    @davout5775 День назад +3

    o1 is a superior model for sure. The price is a big factor tho

    • @YJxAI
      @YJxAI  День назад

      Price is the only retarding factor.

  • @peterweber79
    @peterweber79 День назад +1

    I had the 32b model (locally) draw an ascii-art blue whale. It was a fun task :D

  • @takcalistir
    @takcalistir 2 дня назад +1

    I use both same time. They are all useful for me. Thanks their inventors.

    • @YJxAI
      @YJxAI  2 дня назад +2

      That is the think to do instead of being fanboys.

  • @layonmedia6358
    @layonmedia6358 3 дня назад +3

    i just subbed, i love how you playfully respond to 01 model so fun. good work.

  • @DLBrawlStars
    @DLBrawlStars 3 дня назад +1

    Nice video 😊

  • @FreqtradeFR
    @FreqtradeFR 3 дня назад +2

    Amazing vidéo !

  • @timchen-i9p
    @timchen-i9p День назад +2

    沒有人能拒絕一個強大且免費的工具

  • @onlyyoucanstopevil9024
    @onlyyoucanstopevil9024 12 часов назад +1

    FREE VS $200/MONTH 💀💀💀

  • @vladhanov1530
    @vladhanov1530 3 дня назад +4

    I recently switched to Deepseek, and it's awesome.

    • @YJxAI
      @YJxAI  3 дня назад

      It truly is

  • @Linkwii64
    @Linkwii64 День назад

    I'm addicted to DeepSeek. Can't stop using it

    • @YJxAI
      @YJxAI  День назад

      went too deep you say

  • @RickySupriyadi
    @RickySupriyadi 3 дня назад +4

    2:07 youtuber amazed with dragon svg
    me : trying to understand the excitement
    3:40 oh my god! youtuber seems so excited with this one
    me : wow this model is AGI able to create a Pokémon svg... wow look at that cute happy face

  • @rahi7339
    @rahi7339 18 часов назад

    You should be comparing it to Claude 3.5. I still believe it’s the best for coding

  • @lucids3769
    @lucids3769 3 дня назад +1

    In the mario game o1 could not have come back without that picture u given it so I think deepseck should have that one too

    • @YJxAI
      @YJxAI  3 дня назад

      yeah that is right. But if I think about it the fact that I could share a image to it and improve it's performance is also a plus point to count on the contrary the deep seek model only understands text through OCR. I know it should have been a text to text comparison but I did for this instance.

  • @AllInfo-9
    @AllInfo-9 21 час назад +1

    All codes are taken for web pages some humans did the code.. AI is copy paste of human information but not greate

  • @sangeetanarendrasingh5416
    @sangeetanarendrasingh5416 3 дня назад +2

    You have to use o1 in a particular way, it is way smarter than DeepSeek or Claude or Gemini but also pretty lazy. If you give it the right prompt however its a beast at completing the task in seconds what other models would take several revisions or would be completely unable to do.

    • @blackmartini7684
      @blackmartini7684 3 дня назад +1

      You can say the same about R1, that better prompts give better outputs. R1 just understands what you want off the bat. O1 blows

  • @TheDiamondHawkOfficial
    @TheDiamondHawkOfficial 4 дня назад +4

    W video like always

    • @YJxAI
      @YJxAI  3 дня назад

      Thanks for the constant trust 🥹

  • @the_proffesional1713
    @the_proffesional1713 3 дня назад +2

    What happened to o1-Preview. The previous version was really much better than the new o1. It can answer a complex math problems

    • @YJxAI
      @YJxAI  3 дня назад +1

      I think I should do a o1-preview vs o1 but api costs 🥲

    • @the_proffesional1713
      @the_proffesional1713 3 дня назад

      Try Gemini 2.0 thinking they have a new version i tried to tested on Simplebench it got 7/10 Crazy improvemennts. This is not click bait. I've already tried it yesterday

  • @24-7gpts
    @24-7gpts 14 часов назад +1

    1:36 💀💀

  • @Singlton
    @Singlton 2 дня назад +1

    China: DeepSeek :Dragon!
    America : GPT : green chicken!

  • @shraykumar6892
    @shraykumar6892 День назад +1

    Qwen coder 2.5 32B

    • @YJxAI
      @YJxAI  День назад

      awesome price to performance especially in the backend but not the best.

  • @v300
    @v300 День назад

    Html is not coding.

    • @YJxAI
      @YJxAI  День назад

      I have tested them for python code generation in my full comparison

  • @MrChughad
    @MrChughad 3 дня назад +4

    Hype and nothing.....
    I have tested it all..... useless....

  • @parstigdeyu6259
    @parstigdeyu6259 3 дня назад +2

    I want to learn from the thinking logic instead of outputs

  • @soumen08
    @soumen08 3 дня назад +4

    Maybe if you wrote in proper English, with commas and full stops. Also its stupid, these requests. Ask something serious.

    • @mckane6989
      @mckane6989 3 дня назад +4

      He did a pretty great job comparing them.

    • @user-ze9tj9yj4t
      @user-ze9tj9yj4t 3 дня назад +8

      What a disrespectful and also idiotic comment. The dragon idea was a great way of gauging these models.

  • @revben
    @revben 2 дня назад +1

    O1 is better, but harder to get the results.

    • @YJxAI
      @YJxAI  2 дня назад

      Spoiled kid need to scold. 🤣

  • @XoloMariduinia
    @XoloMariduinia 3 дня назад +2

    🎉

  • @kuchbhi10507
    @kuchbhi10507 3 дня назад +2

    Both r createed by same company openai