Elon Musk’s updated Grok AI claims to be better at coding and math

grok ai
0
0

Elon Musk’s xAI, aimed at rivaling ChatGPT, is undergoing an update to enhance its proficiency in math, coding, and more. Grok-1.5, the latest iteration, has been rolled out to early testers, boasting “improved capabilities and reasoning,” along with the ability to process longer contexts. According to xAI’s claims, Grok-1.5 demonstrates significant advancements over its predecessor, Grok-1, achieving a noteworthy 50.6 percent in the MATH benchmark, more than double its previous score. Furthermore, it has elevated its performance to 90 percent and 74.1 percent in GSM8K (math word problems) and HumanEval (coding), respectively, compared to 62.9 percent and 63.2 percent previously. These figures bring Grok-1.5 into close competition with Gemini Pro 1.5, GPT-4, and Claude 3 Opus, with its HumanEval coding score surpassing all rivals except Claude 3 Opus.

Moreover, Grok-1.5 exhibits the capability to process long contexts of up to 128K tokens within its context window, enabling it to amalgamate data from numerous sources to comprehend complex situations. “This allows Grok to have an increased memory capacity of up to 16 times the previous context length, enabling it to utilize information from substantially longer documents,” stated the company.

While xAI hasn’t elaborated on Grok’s advancements in other areas, such as academic scores and multimodal capabilities, Grok-1.5’s reign may be short-lived as ChatGPT 5 is slated to launch this summer, promising an enhanced feature set that aims to simulate human-like communication.

Presently, Grok is exclusively accessible to Premium+ tier users on X (previously Twitter), although Elon Musk has recently pledged to extend access to regular Premium users on the platform. Additionally, the company has recently made its Grok chatbot open source, following Musk’s lawsuit against OpenAI and Sam Altman for allegedly deviating from its non-profit mission.