Last year, Twitter, now known as X, launched its chatbot Grok-1.5 designed for coding and mathematics tasks. Initially only available to Premium+ users, the AI model has since been expanded to all Premium subscribers. During performance tests, Grok-1.5 scored well in math benchmark tests, achieving a 50.6% score in a comparative math test and a 90% score in GSM8K. These tests covered a wide range of competency problems from elementary school to high school levels.
One of the key improvements in the new iteration of the chatbot is its increased ability to process long contexts of up to 128,000 tokens within its context window. This allows Grok-1.5 to have an expanded memory capacity, enabling it to use information from much longer documents than before.
X has announced that Grok-1.5 will be available to some users next week, with the aim of receiving feedback to further improve the tool. The company also plans to introduce several new features in the coming days as it reaches a broader audience. Overall, the new version of X’s chatbot is designed to enhance its abilities in coding and mathematics tasks with improved reasoning capabilities and an increased memory capacity for processing longer contexts.