Elon Musk’s Grok AI is now available in India: More expensive than ChatGPT Plus

Grok’s generative language model offers two modes—fun and regular. As the name suggests, the fun mode offers responses that are more jovial in nature, while the regular mode offers standard responses. The biggest advantage of Grok AI over other language models is its ability to access real-time data from X, enabling it to construct more accurate answers even on recent trends and developments.

xAI has established that, despite being a model with 63.2 billion parameters, Grok-1 is more capable than GPT-3.5, which powers the free version of ChatGPT. The company also claims that Grok-0, the previous model with just 33 billion parameters, can outperform Meta’s LLaMA 2 with 70 billion parameters on various tasks like HumanEval coding and middle-class math word problem-solving.

In November, several users highlighted that Grok AI includes phrases like ‘cannot fulfil the request, as it goes against OpenAI’s use case policy,’ indicating the possible use of OpenAI’s code. However, Igor Babuschkin, the chief engineer of xAI, quickly came to the rescue of Grok and stated that the model has been trained on a vast amount of internet data, which also contains data generated by GPT. He noted that it is a very rare issue and assured that no OpenAI code was used in developing Grok.