Tell me the latest news on the llama 4 launch, which of the models are available on aws…
- User
Tell me the latest news on the llama 4 launch, which of the models are available on aws , and in what form
Assistant
Here's the latest information regarding the Llama 4 launch and its availability on AWS:
Llama 4 Launch and Models
- Meta has released the first two models in its Llama 4 series: Llama 4 Scout and Llama 4 Maverick.[[1]][[3]][[4]]
- Llama 4 relies on a novel design that combines a Mixture of Experts architecture with an early-fusion backbone that allows it to be natively multimodal.[[2]][[4]][[6]]
- Llama 4 Scout (109B total parameters, 17B active parameters) with 16 experts.[[2]][[4]]
- Llama 4 Maverick (400B total parameters, 17B active parameters) with 128 experts.[[2]][[4]]
- Llama 4 Scout is designed for tasks like multi-document summarization, parsing user activity for personalized tasks, and reasoning over vast codebases, and supports a 10M token context window.[[4]]
- Llama 4 Maverick excels at image and text understanding for general assistant and chat use cases.[[3]]
- Meta is also developing Llama 4 Behemoth, a model with 288 billion active parameters.[[3]][[5]]
- Llama 4 Reasoning is the fourth model and more information will be provided about that model soon.[[3]][[6]]
Availability on AWS[[1]][[7]][[11]]
- Llama 4 Scout 17B and Llama 4 Maverick 17B are available on Amazon SageMaker JumpStart.[[1]]
- Availability as a fully managed, serverless option in Amazon Bedrock is coming soon.[[1]]
- Llama 3 models are also available on AWS.[[7]][[9]][[12]]
- The Llama 3 models are available in Amazon Bedrock and Amazon SageMaker.[[6]][[10]] They are also available via Amazon Elastic Compute Cloud (Amazon EC2) using AWS Trainium and Inferentia.[[6]][[7]][[10]]
Model Details and Capabilities
- Multimodality: Llama 4 models are built with native multimodality, handling both text and images.[[1]][[4]][[13]]
- Context Window: Llama 4 Scout supports a context window of up to 10 million tokens.[[2]][[3]][[5]]
- Efficiency: The models use a Mixture of Experts (MoE) architecture.[[1]][[4]][[5]]
- Languages: Some models support multilingual dialogue use cases across English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.[[7]][[10]][[14]]
Llama 3 Models on AWS[[6]][[11]][[12]]
- Llama 3.3 70B is available in Amazon Bedrock and Amazon SageMaker AI, and via Amazon Elastic Compute Cloud (Amazon EC2).[[10]]
- Llama 3.2 models are available on Amazon Web Services (AWS) in Amazon Bedrock and Amazon SageMaker AI.[[6]][[10]] They are also available via Amazon Elastic Compute Cloud (Amazon EC2) using AWS Trainium and Inferentia.[[6]][[7]][[10]]
- Llama 3 is available via Amazon SageMaker JumpStart.[[6]][[8]][[11]] Llama 3 comes in a range of parameter sizes — 8B and 70B.[[10]][[11]][[14]]
- Meta Llama 3 models are available in Amazon Bedrock in the US East (N. Virginia) and US West (Oregon) Regions.[[6]][[9]]
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.