Amazon Bedrock Announces Open Weight Models for AI Coding Agents

Amazon Bedrock has announced the availability of open weight models for AI coding agents, providing developers with a secure and cost-effective way to leverage AI in their coding processes. According to the announcement, these models allow developers to run AI coding agents that keep proprietary data within their own AWS accounts, switch between different models on demand, and charge only for actual usage.

The new feature is supported by OpenCode, an open-source, terminal-native AI coding agent built in Go. OpenCode reads and edits files, runs shell commands, and understands project structure through Language Server Protocol (LSP) diagnostics. It connects to over 75 large language model (LLM) providers, including Amazon Bedrock, and allows developers to set up multi-model workflows that match the right model to each task.

In the announcement, Amazon Bedrock demonstrates how to set up OpenCode with open weight models on Amazon Bedrock, configure multi-model workflows, and provides practical coding examples using models such as Moonshot AI Kimi K3, OpenAI GPT-OSS 120B, and NVIDIA Nemotron 3 Super 120B. The setup process includes installing the OpenCode CLI and configuring AWS credentials.

This new capability offers developers the flexibility to choose from a wide range of models and the security of running them within their AWS environment, eliminating the need for per-seat fees and third-party APIs. The announcement highlights the potential for cost efficiency and customization, along with enterprise security and managed infrastructure provided by Amazon Bedrock.

Source: aws-ml