XAI founder testing Grok 3 model, launch imminent

Generative AI company, xAI Founder Elon Musk is testing an early version of Grok 3 large language model (LLM) chatbot for an upcoming release.

“Testing Grok 3 int4 Inference,” wrote Musk while confirming this new model launch preparations on social media site X (formerly Twitter).

Grok is an LLM launched in December 2023. So far, xAI seeded two major releases: v1.5 and v2.0. These two have increased the response rate, accuracy, and brand-new features.

Especially, Grok 2 has new image generation powered by Auroroa image generator, file analysis, web search, and more.

LLM inference is the mechanism large language models use to generate human-like responses. Once a generative model receives an input prompt from the user, it uses knowledge gained during training to predict the most likely tokens in the sequence before decoding those tokens into text outputs.

Meanwhile, INT4 offers computational speedups with a reduction a network’s memory footprint as well as conserve memory bandwidth to ease to run multiple ensemble networks on a single GPU.

Musk also shared a screenshot of the interference test including a question and a response showing the vibrance in color related to the intent of the user query.

Credit – @elonmusk/X

Musk previously said that the Grok 3 would have 10 times more than 10 times computing power than the Grok 2. However, a launch date for this new model is not yet announced.