Elon Musk teases next-gen AI chatbot Grok-1.5 with superior coding and math skills
This revelation came via Musk’s social media post on March 29, following xAI’s announcement of Grok-1.5 in a blog post. The enhanced AI chatbot will initially be accessible to “early testers and existing Grok users on the social media platform.”
Additionally, Musk hinted at the ongoing development of Grok 2, which he anticipates will surpass current AI standards in all aspects.
Grok-1.5
Grok-1.5 is an advanced version of the Grok-1 AI model and comes with improved reasoning and a context length of 128,000 tokens.
xAI’s analysis indicates significant enhancements in the performance of its advanced chatbot, particularly in coding and math-related tasks. However, it falls short compared to Google’s Gemini Pro 1.5 and OpenAI’s GPT-4.
According to the firm:
“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.”
Moreover, Grok-1.5 can utilize information from substantially longer documents, and the model can handle longer and more complex prompts while maintaining its instruction-following capability as its context window expands.
The firm 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 prototype ideas and train new architectures at scale with minimal effort.”
Grok is Open-source
Earlier this month, xAI took a significant step by open-sourcing the base code of Grok-1.
This decision arose as a response to a legal action initiated by Musk against OpenAI, the organization he once co-founded. Musk alleged that OpenAI has deviated from its original commitment to prioritize open-source model development over shareholder interests.
Meanwhile, xAI said the released code was “the raw base model checkpoint from the Grok-1 pre-training phase, which concluded in October 2023. This means that the model is not fine-tuned for any specific application, such as dialogue.” It added that the model was licensed under Apache License 2.0.
