The Mistral Large AI Model stands out as a remarkable achievement. Developed by Mistral AI and first available on Azure AI Studio, this model is designed to deliver on any text-based use case, thanks to its state-of-the-art reasoning and knowledge capabilities. This article will guide you through the process of deploying and exploring the Mistral Large AI Model in Azure AI Studio, a user-friendly interface that allows you to interact with the model and explore its capabilities.
Mistral Large AI Model in Azure AI Studio
Microsoft and Mistral AI have announced a multi-year partnership to accelerate AI innovation and introduce the Mistral Large model first on Azure. This partnership enables Mistral AI with access to Azure’s cutting-edge AI infrastructure, accelerating the development and deployment of their next-generation large language models (LLMs).
The Mistral Large model, Mistral AI’s flagship commercial model, is now available first on Azure AI and the Mistral AI platform. This marks a noteworthy expansion of their offerings. Mistral Large is a general-purpose language model that can deliver on any text-based use case thanks to its state-of-the-art reasoning and knowledge capabilities.
Microsoft will support Mistral AI with Azure AI supercomputing infrastructure, delivering best-in-class performance and scale for AI training and inference workloads for Mistral AI’s flagship models. Microsoft and Mistral AI will make Mistral AI’s premium models available to customers through the Models as a Service (MaaS) in the Azure AI Studio and Azure Machine Learning model catalog.
This partnership represents an opportunity for Mistral AI to unlock new commercial opportunities, expand to global markets, and foster ongoing research collaboration. It’s a significant milestone in the AI industry and a testament to the impactful progress driven by both Microsoft and Mistral AI.
What is the Mistral Large Model?
Mistral Large is a model offered by Mistral AI in Azure AI Studio. It is Mistral AI’s most advanced Large Language Model (LLM) and can be used on any language-based task thanks to its state-of-the-art reasoning and knowledge capabilities.
Key features of Mistral Large include:
Specialization in RAG (Retrieval-Augmented Generation), ensuring crucial information isn’t lost in the middle of long context windows (up to 32 K tokens).
Strong in coding, supporting code generation, review, and comments in all mainstream coding languages.
Multilingual by design, offering best-in-class performance in French, German, Spanish, and Italian, in addition to English. Dozens of other languages are also supported.
Responsible AI, with efficient guardrails baked into the model, and an additional safety layer with the safe_mode option.
In Azure AI Studio, you can deploy the Mistral Large model as a service with pay-as-you-go billing. This deployment option doesn’t require quota from your subscription. Please note that Mistral AI can change or update the terms of use and pricing of this model.
The cost of using the Mistral Large model in Azure AI Studio is based on the number of tokens for both input and output.
The cost is approximately $0.024 per 1000 tokens for output
The cost is $0.008 per 1000 tokens for input.
You can review the pricing on the Mistral Large offer in the Marketplace offer details tab when deploying the model or on the Azure Marketplace.
Quota
In Azure AI Studio, the quota is handled on a per-deployment basis. Each deployment has a rate limit of 200,000 tokens per minute and 1,000 API requests per minute.
However, currently, there is a limit of one deployment per model per project. If the current rate limits do not meet your requirements, please contact Microsoft Azure Support for assistance.
Step by steps guide to Deploy Mistral Large Model
STEP 1: Open Azure AI Studio. Navigate to “Explore => Model Catalog”.
STEP 2: Search for Mistral-large. To find the Mistral-large models, click on "View Models" under the Mistral model. Then click on that model.
Alternatively, you can also go to “Deployment”. Click on “+ Create” and select the "Pay-as-you-go" option.
After that, select the Mistral-large model and click on "Select".
STEP 3: In the model catalog, click on “Deploy”. Then select the “Pay-as-you-go” option.
STEP 4: Choose the project in which you want to deploy your model. In Azure AI Studio, to deploy the Mistral-large model, your project must be in the East US 2 or France Central regions.
To find out the pricing of the Mistral Large model based on tokens, navigate to the “Marketplace Offer Details” section.
STEP 5: Select “Subscribe and Deploy”.
If this is your first time deploying the model in the project, you have to subscribe to your project for the particular offering. This step requires that your account has the Azure AI Developer role permissions on the Resource Group, as listed in the prerequisites.
Each project has its own subscription to the particular Azure Marketplace offering of the model, which allows you to control and monitor spending.
Currently, you can have only one deployment for each model within a project.
STEP 6: Give the deployment a name. This name becomes part of the deployment API URL. This URL must be unique in each Azure region.
STEP 7: Click on “Deploy”. Wait until the deployment is ready and you’re redirected to the “Deployments” page.
STEP 8: You can note the endpoint’s “Target URL” and the “Secret Key”, which you can use to call the deployment for chat completions using the <target_url>/v1/chat/completions API.
Mistral-Large Playground
Once your model is deployed, you can utilize it for your specific tasks. To do this, click on “Open in playground”.
Alternatively, you can navigate to the “Playground” and then select your model.
In Azure AI Studio, there are three modes available (Until now):
Chat
Completion
Images
Chat: The chat functionality allows you to interact with the model in a conversational manner. You can input a series of messages, and the model will respond accordingly. This is useful for testing the model’s ability to maintain context over a conversation, answer questions, and generate creative content.
Suppose you want to test the model’s conversational abilities. You can enter a series of messages in the chat interface and observe how the model responds. For example, you could start a conversation with “What is a Time and weather in Toronto?” and continue the conversation based on the model’s responses.
Completion: The completion functionality is used to generate text that completes a given prompt. You provide an initial piece of text, and the model generates a continuation of that text. This is useful for a wide range of tasks, such as writing assistance, idea generation, and more.
If you’re interested in text generation, you can use the completion mode. For instance, you could enter a prompt like “Once upon a time,” and see how the model completes the story.
Image: While the Mistral Large model is primarily a text-based model, it can work with image data in certain contexts. For example, you could upload an image and ask the model to generate a description or a story based on the image.
Conclusion
The Mistral Large AI Model in Azure AI Studio offers a powerful tool for a wide range of applications, from virtual assistants to content creation. Its versatility and advanced capabilities make it a valuable asset for anyone looking to leverage the power of AI.