Elon Musk previews Grok-1.5, the next generation of AI chatbot, with better programming and math skills

Grok-1.5 is an advanced version of the Grok-1 AI model and brings improved reasoning capabilities.

Elon Musk has announced that an upgraded version of his artificial intelligence company xAI’s chatbot Grok may be released next week.

The news was released via Musk’s social media on March 29, referring to the announcement of Grok-1.5’s AI in a blog post. The enhanced AI chatbot will be available to "early testers and existing Grok users on social media platforms."

Additionally, Musk hinted at the continued development of Grok 2, a product he expects to surpass current AI standards in all aspects.

Grok-1.5

Grok-1.5 is an advanced version of the Grok-1 AI model with improved inference capabilities and a context length of 128,000 tokens.

xAI’s analysis shows that its advanced chatbot has significantly improved performance, especially in programming and math-related tasks. It's still slightly inferior compared to Google's Gemini Pro 1.5 and OpenAI's GPT-4.

Analysis of AI Chatbots | Source: xAI

According to the company:

"Grok-1.5 achieved a score of 50.6% on the MATH benchmark and a score of 90% on the GSM8K benchmark, both mathematics benchmarks covering a wide range of competition problems from elementary school to high school. Furthermore, it achieved It scored 74.1% on the HumanEval benchmark, which evaluates code generation and problem-solving abilities."

Additionally, Grok-1.5 can leverage information from longer documents, and the model can handle longer, more complex prompts while maintaining its instruction tracking capabilities as the context window expands.

The company added:

"Grok-1.5 is built on a custom distributed training framework based on JAX, Rust, and Kubernetes. This training stack enables our team to build ideas and train new architectures at scale with minimal effort."

Grok is open source

Earlier this month, xAI took a major step forward by opening up the underlying code of Grok-1.

The decision comes in response to Musk's legal action against the OpenAI organization he once co-founded. . Musk claims that OpenAI has strayed from its original commitment to prioritize developing open source models over shareholder interests.

Meanwhile, xAI says the released code is "the original base model checkpoint from the Grok-1 pre-training phase, which ended in October 2023. This means that the model has not been fine-tuned for any specific application (such as dialogue). ” It added that the model is licensed under the Apache License 2.0.

The above is the detailed content of Elon Musk previews Grok-1.5, the next generation of AI chatbot, with better programming and math skills. For more information, please follow other related articles on the PHP Chinese website!