Meta’s new Llama 3.3 70B Instruct model now available on watsonx.ai

IBM announces the availability of the updated Llama 3.3 70B Instruct model on IBM watsonx.ai™. The updated Llama model offers performance comparable to the popular Llama 3.1 405B model, at a significantly lower cost. This makes it an ideal choice for developers seeking optimal performance without breaking the bank.

The updated Llama 3.3 70B Instruct model performs better at coding, reasoning, tool use and general knowledge tasks. The model’s enhanced capabilities include the ability to generate JSON outputs for function calling, provide step-by-step reasoning with accurate responses, and it also offers greater coverage and accuracy for a wide range of programming languages. It also delivers improved code feedback, error handling and fixing, and can invoke relevant tools as and when required. The model will continue to offer multilingual support for various languages including English, German, French, Italian, Portuguese, Hindi, Spanish and Thai.

The release of the updated Llama 3.3 Instruct 70B model on IBM’s watsonx.ai model catalogue continues to build on the work initiated by IBM and Meta last year, aimed at fostering open innovation in AI. Watsonx.ai already supports the Llama 3.2 models as well. This collaboration unites leading organizations, startups, researchers, and academia through the AI Alliance to promote responsible and open AI development globally.

Companies are increasingly using Llama on watsonx.ai as they turn to open AI models and tools to create clear business value. Watsonx.ai offers a modern enterprise studio where AI developers can train, test, fine-tune, and deploy various models, including Llama. Businesses in various sectors, from telecom and financial services to local governments, are benefiting from watsonx.ai and Llama models.

Developers can now inference the updated Llama 3.3 70B Instruct model on watsonx.ai.

IBM watsonx.ai is an enterprise-grade AI studio for builders—from the application developer to the data scientist and everyday line-of-business user. As a one-stop AI developer studio, watsonx.ai combines gen AI with traditional ML techniques to help builders innovate with all the APIs, models, tools and runtimes to simplify and scale the development and deployment of AI applications.

Within watsonx.ai, AI builders can support the rapid adoption of AI use cases, from data through deployment, by using a collection of foundation models, including Llama. With watxonx.ai, uses can leverage an enterprise-grade developer AI toolkit, manage the full AI lifecycle and bring traditional AI and ML into production faster.

Start using watsonx.ai today for free.