Tech giant Microsoft announced it is adding Elon Musk’s AI startup xAI’s Grok 3 to its Azure platform.
Grok 3 and Grok 3 mini will be available through Azure’s AI Foundry platform, which allows developers to design, manage, customize and support enterprise-ready AI agents and apps.
Grok 3, xAI’s flagship model, comprises LLMs focused on pushing AI innovation and accelerating scientific discovery. The companies said the model was trained on “xAI’s Colossus supercluster with 10x the compute power of prior leading models.”
Microsoft said in a statement that the model excels in mathematics, instruction-following, coding, reasoning and world knowledge and has deep domain expertise in numerous sectors, including healthcare and science. Grok and Microsoft tout that the model has medical diagnosis support and can deliver scientific research assistance.
Grok will be available through a free preview for two weeks, and subsequently with an inputting price of $3 for one million tokens for Grok 3 (Global) and $3.30 for Grok 3 (DataZone), and an outputting price of $15 for one million tokens for Grok 3 (Global) and $16.50 for Grok 3 (DataZone). The model is also accessible on GitHub.
Microsoft’s Foundry platform hosts other AI models, including those from NVIDIA, Meta, Cohere and OpenAI.
“This collaboration combines xAI’s cutting-edge models with Azure’s enterprise-ready infrastructure, giving developers access to Grok 3’s advanced capabilities in a secure, scalable environment. Grok models enable a range of enterprise scenarios with advanced capabilities in reasoning, coding and visual processing,” Microsoft said in a statement.
THE LARGER TREND
In November, Musk posted on X that users should “try submitting X-ray, PET, MRI or other medical images to Grok for analysis. This is still early stage, but it is already quite accurate and will become extremely good.”
The billionaire tech mogul said users of the chatbot could upload their medical images and generate a diagnosis. He then requested users to let the AI maker know where Grok “gets it right or needs work.”
Skeptics have relayed their concerns around the AI model generating diagnoses, and European privacy regulators have questioned potential violations of the General Data Protection Regulation (GDPR) due to the way Grok processes data.
On April 11, the Irish Data Protection Commission (DPC), Ireland’s privacy watchdog, launched an inquiry into how X uses Europeans’ personal data to train Grok, specifically investigating whether Musk’s platform uses publicly-accessible posts made on X to train the genAI model. The inquiry will determine compliance with GDPR rules.
xAI started laying the foundation for its “anti-woke” Grok-1 chatbot in November 2023.
“Grok is designed to answer questions with a bit of wit and has a rebellious streak, so please don’t use it if you hate humor,” xAI said in Grok’s launch announcement.
