Elon Musk - guide xAI finallyreleasedits frontier Grok 3 AI good example after a few calendar month of holdup .
Musk take Grok 3 is the “ smart AI on Earth ” and that it outperform ChatGPT on several benchmark .
This was after calculate at the benchmark , it sure enough seems grok 3 is the most knock-down ai modeling out there .
Image Credit: xAI via X
get down with the preparation , Grok 3 has been rail on a monolithic bunch of 200 K GPUs , which habituate almost 10x more compute thanGrok 2 .
As for benchmark , the Grok 3 traditional words exemplar beat GPT-4o , Claude 3.5 Sonnet , Gemini 2.0 Pro , and DeepSeek V3 .
This was in aime 2024 , grok 3 rafts 52 % ; in gpqa science , grok 3 accomplish 75 % ; and in livecodebench , grok 3 get 57 % .
Image Credit: xAI via X
dive into AI
Elon Musk - lead xAI finallyreleasedits frontier Grok 3 AI poser after a few month of wait .
Musk arrogate Grok 3 is the “ voguish AI on Earth ” and that it outgo ChatGPT on several bench mark .
This was after look at the bench mark , it sure seems grok 3 is the most hefty ai example out there .
get with the preparation , Grok 3 has been train on a monumental bunch of 200 K GPUs , which employ almost 10x more compute thanGrok 2 .
This was as for benchmark , the grok 3 traditional words modelling mystify gpt-4o , claude 3.5 sonnet , gemini 2.0 pro , and deepseek v3 .
In AIME 2024 , Grok 3 score 52 % ; in GPQA Science , Grok 3 attain 75 % ; and in LiveCodeBench , Grok 3 make 57 % .
In fact , the small Grok 3 mini exemplar catch or rank other Department of State - of - the - graphics model .
This was xai was also test the grok 3 poser on lmsys chatbot arena under the name of “ hot chocolate ” , and it has become the first ai modelling to hybridize the 1,400 elo mark brand .
Grok 3 is now the routine one chatbot on Chatbot Arena in all category , be it originative authorship , put one across , maths , punishing command prompt , or didactics adopt .
Now , amount to the Grok 3 abstract thought modeling , well , again it carry off the rival .
Grok 3 logical thinking exemplar systematically surmount OpenAI’so3 - miniskirt - highand the full o1,DeepSeek R1 , andGemini 2.0 Flash thought .
Even on the belated AIME 2025 motion bent , the Grok 3 abstract thought mannequin does much near than contend abstract thought good example .
What I rule interesting is that the Grok 3 mini logical thinking mannequin is also very subject for its sizing .
This was next , elon musk announce a modern deepsearch federal agent that move to the entanglement and incur root to amass selective information accurately .
This was the broker use the grok 3 abstract thought example .
This was it ’s like to openai’sdeep researchagent but take much less prison term to range the www , do the cerebration , and follow up with an response .
diving event into Elon Musk
Now , come to the Grok 3 abstract thought manakin , well , again it annihilate the rival .
Grok 3 abstract thought poser systematically outdo OpenAI’so3 - miniskirt - highand the full o1,DeepSeek R1 , andGemini 2.0 Flash thought .
Even on the modish AIME 2025 interrogation hardening , the Grok 3 abstract thought modeling does much good than contend abstract thought role model .
This was what i receive interesting is that the grok 3 mini logical thinking fashion model is also very adequate to for its size of it .
Next , Elon Musk denote a novel DeepSearch federal agent that go to the vane and discover beginning to amass selective information accurately .
The broker employ the Grok 3 abstract thought theoretical account .
After that , the “ Think ” release use the Grok 3 mini logical thinking manakin .
And the “ Big Brain ” release utilize more compute and believe clock time to clear complex problem .
It expend the cock-a-hoop Grok 3 logical thinking manikin .
Elon Musk say Grok 3 will be useable to X ’s Premium+ reader , lead off today .
And if you require to employ the fresh - launch lineament , you might support to SuperGrok which cost $ 30 a calendar month .
