Elon Musk's X to introduce advanced Grok-1.5 AI to testers next week

Share this article

Grok-1.5, the latest artificial intelligence (AI) model developed by Elon Musk’s AI division xAI, will be available for early testers and Grok users on X next week, according to Musk’s recent post.

Should be available on 𝕏 next week.

Grok 2 should exceed current AI on all metrics. In training now. https://t.co/s7rVxvK8mZ

— Elon Musk (@elonmusk) March 29, 2024

The latest version has shown impressive improvements compared to Grok-1, xAI said in a blog post on Thursday, including enhanced reasoning and problem-solving and an extended context window while maintaining instruction-following.

According to xAI, Grok-1.5 excels in coding and math tasks, achieving high scores on established benchmarks.

“Grok-1.5 achieved a 50.6% score on the MATH benchmark and a 90% score on the GSM8K benchmark, two math benchmarks covering a wide range of grade school to high school competition problems. Additionally, it scored 74.1% on the HumanEval benchmark, which evaluates code generation and problem-solving abilities,” stated xAI.

This improvement allows the tool to handle complex prompts while maintaining its ability to follow instructions. Grok-1.5 can now process extended contexts up to 128,000 tokens, allowing for a better understanding of complex information.

The team will make Grok-1.5 available for early access on X next week, with further features planned for a wider rollout.

xAI open-sources Grok-1

xAi has recently made the core code of Grok-1 publicly available. This move comes amid a legal battle between Musk and OpenAI, another AI research organization he co-founded. Musk claims OpenAI has strayed from its initial goal of prioritizing open-source AI models, focusing more on shareholder interests.

xAI clarifies that the released code is the “raw base model checkpoint” from Grok-1’s pre-training phase. This means it’s not yet tailored to specific tasks like conversation.

Share this article