Join our daily and weekly newsletters to stay updated with the latest and exclusive content on leading AI industry developments.
Microsoft has been a big player in the generative AI boom, largely due to its early investment in OpenAI and the integration of OpenAI’s technology into Bing, Azure, and various other services. However, Microsoft isn’t putting all its AI investments in one place.
Today, Microsoft announced that it will offer Meta Platforms’ Llama 2, an open-source AI model, in Azure AI Studio as a “model-as-a-service.”
Understanding models-as-a-service
Similar to Software-as-a-Service (SaaS), Models-as-a-Service (MaaS) allows customers to use AI models like Llama 2 on demand through the web with minimal setup. This means they won’t need to go through the hassle of installing it on their own servers or cloud spaces, saving significant time and effort even with expert help.
John Montgomery, Microsoft’s Corporate VP of Program Management for its AI Platform, explained in a blog post that Azure AI has long allowed customers to deploy models on their own infrastructure. But not all customers want to manage this, so Microsoft introduced Models-as-a-Service. This new service operates models as API endpoints that can be easily called upon, much like the Azure OpenAI Service. Now, Llama 2 models, including Llama-2-7b (Text Generation), Llama-2-7b-Chat (Chat Completion), Llama-2-13b (Text Generation), Llama-2-13b-Chat (Chat Completion), Llama-2-70b (Text Generation), and Llama-2-70b-Chat (Chat Completion), are available in public preview through Azure AI.
Offering these open-source Llama models is a strategic move for Microsoft. It provides Azure cloud customers with more AI choices and a lower-cost alternative to OpenAI’s GPT-3.5 and 4 models. Llama 2 has become a favorite among many users and enterprises for generative AI.
There were reports that Microsoft CEO Satya Nadella was surprised and upset by the former OpenAI board’s decision to terminate CEO Sam Altman, leading to his reinstatement. This has prompted Nadella to diversify Microsoft’s AI investments.
Despite these changes, OpenAI remains a significant part of Microsoft’s strategy. Montgomery announced that OpenAI’s latest AI model, GPT-4 Turbo with Vision, which can analyze and describe photos and visual material, is now available to Azure customers. Companies like Instacart and WPP are already using this model.
Additionally, the Azure AI Studio includes tools for fine-tuning all the models available. As the competition in the AI cloud market continues, the focus is on how many models Microsoft can integrate into Azure AI Studio next.
Recently, Microsoft also released Phi-2, a small language model intended for research purposes, not commercial use. However, Llama-2 and GPT-4 Turbo with Vision are available for commercial applications.
Stay informed with our latest news by subscribing to our daily updates.