Introducing Reflection Beam: an open-weight model for agentic coding on AWS
Reflection AI's Beam is an open-weight model for agentic coding and long-horizon tasks. This post shows how to deploy Beam on AWS: start with an Amazon SageMaker JumpStart real-time endpoint, or use Amazon SageMaker HyperPod Inference with JumpStart, SGLang, or vLLM for Kubernetes-native control over distributed GPU infrastructure.
Reflection AI has unveiled Beam, an open-weight model designed for agentic coding and complex, multi-step tasks like coding, tool usage, result interpretation, and workflow completion. Beam boasts an impressive 501 billion total parameters and 23 billion active parameters, with a context length of 1 million tokens. The model weights are made available on Hugging Face under the Apache 2.0 license.
Developers and researchers can deploy Beam using AWS resources. For a fully managed solution, an Amazon SageMaker JumpStart real-time endpoint is recommended. Alternatively, Amazon SageMaker HyperPod Inference with JumpStart, SGLang, or vLLM offers Kubernetes-native control over distributed GPU infrastructure. Beam’s parameters influence checkpoint storage, weight memory, sharding, parallelism, and per-token compute.
Currently available in US East (N. Virginia) and US East (Ohio), Beam follows Reflection’s vision of open intelligence, aiming to make AI accessible to everyone through open model weights, research, and software. The model is available in BF16, FP8, and NVFP4 checkpoint variants, with deployment options through SageMaker JumpStart, Hugging Face Hub, or an optional Amazon S3 mirror.
Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
