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

It’s Elon Musk Reply to ChatGPT it’s getting updates to improve math, writing and more. Musk’s xAI has been launched Price – 1.5 for early adopters with “excessive ability and imagination” and the ability to make long-term changes. The company says it is now working on GPT-4, Gemini Pro 1.5 and Close 3 Functions in several areas.

Going by the xAI numbers, the Grok-1.5 looks like a big improvement over the Grok-1. It rose to 50.6 percent in the MATH benchmark, double the previous record. It also rose 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 numbers are within shouting distance of Gemini Pro 1.5, GPT-4 and Claude 3 Opus – in fact, HumanEval codes beat all competitors except Claude 3 Opus.

Elon Musk's latest Grok AI boosts writing and math skills


It can also process up to 128K tokens within its window, meaning it can integrate information from multiple sources to get a better understanding of the situation. “This allows Grok to have the ability to remember information up to 16 times the length of previous articles, making it more efficient to use information from long documents,” the company said.

xAI did not explain in detail how Grok is going in other areas, however, where it may be left behind (a lot of education, multimodal and others). And Grok-1.5 will not hold its position for long. ChatGPT 5 is expected to arrive this summer, promise a feature that “makes it feel like you’re interacting with a human rather than a machine,” according to OpenAI.

Currently, Grok is available to Premium+ tier users on X (formerly Twitter), even Elon Musk recently. he promised to unlock for regular X Premium users. The company recently open sourced his Grok chatbot, after Musk opposed OpenAI and Sam Altman for leaving his nonprofit.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *