Llama models are available for use as managed APIs and self-deployed models on Vertex AI. You can stream your responses to reduce the end-user latency perception. A streamed response uses server-sent events (SSE) to incrementally stream the response.
Managed Llama models
Llama models offer fully managed and serverless models as APIs. To use a Llama model on Vertex AI, send a request directly to the Vertex AI API endpoint. When using Llama models as a managed API, there's no need to provision or manage infrastructure.
The following models are available from Llama to use in Vertex AI. To access a Llama model, go to its Model Garden model card.
Llama 4 Maverick 17B-128E
Llama 4 Maverick 17B-128E is the largest and most capable Llama 4 model that offers coding, reasoning, and image capabilities. It features Mixture-of-Experts (MoE) architecture with 17 billion active parameters out of 400 billion total parameters and 128 experts. Llama 4 Maverick 17B-128E uses alternating dense and MoE layers, where each token activates a shared expert plus one of the 128 routed experts. The model is pretrained on 200 languages and optimized for high-quality chat interactions through a refined post-training pipeline.
Llama 4 Maverick 17B-128E is multimodal and is suited for advanced image captioning, analysis, precise image understanding, visual questions and answers, creative text generation, general-purpose AI assistants, and sophisticated chatbots requiring top-tier intelligence and image understanding.
Considerations
- You can include a maximum of three images per request.
- The MaaS endpoint doesn't use Llama Guard, unlike previous versions. To use Llama Guard, deploy Llama Guard from Model Garden and then send the prompts and responses to that endpoint. However, compared to Llama 4, Llama Guard has a more limited context (128,000) and can only process requests with a single image at the beginning of the prompt.
- Batch predictions aren't supported.
Llama 4 Scout 17B-16E
Llama 4 Scout 17B-16E delivers state-of-the-art results for its size class that outperforms previous Llama generations and other open and proprietary models on several benchmarks. It features MoE architecture with 17 billion active parameters out of the 109 billion total parameters and 16 experts.
Llama 4 Scout 17B-16E is suited for retrieval tasks within long contexts and tasks that demand reasoning over large amounts of information, such as summarizing multiple large documents, analyzing extensive user interaction logs for personalization, and reasoning across large codebases.
Considerations
- You can include a maximum of three images per request.
- The MaaS endpoint doesn't use Llama Guard, unlike previous versions. To use Llama Guard, deploy Llama Guard from Model Garden and then send the prompts and responses to that endpoint. However, compared to Llama 4, Llama Guard has a more limited context (128,000) and can only process requests with a single image at the beginning of the prompt.
- Batch predictions aren't supported.
Llama 3.3
Llama 3.3 is a text-only 70B instruction-tuned model that provides enhanced performance relative to Llama 3.1 70B and to Llama 3.2 90B when used for text-only applications.
Go to the Llama 3.3 70B model card
Use Llama models
For managed models, you can use curl commands to send requests to the Vertex AI endpoint using the following model names. To learn how to make streaming and non-streaming calls to Llama models, see Call open model APIs.
To use a self-deployed Vertex AI model:
- Navigate to the Model Garden console.
- Find the relevant Vertex AI model.
- Click Enable and complete the provided form to get the necessary commercial use licenses.
For more information about deploying and using partner models, see Deploy a partner model and make prediction requests.
What's next
Learn how to use Llama models.