OpenCode pairs with Bedrock-hosted open weight models to keep code and inference inside a customer’s own AWS account.

AWS detailed how developers can run OpenCode, an open source, terminal-native coding agent built in Go, against open weight models hosted on Amazon Bedrock. OpenCode reads and edits files, runs shell commands, and uses Language Server Protocol diagnostics to understand project structure, while inference stays inside the customer’s AWS account.

OpenCode connects to more than 75 LLM providers, including Bedrock, and the post frames open weight adoption with McKinsey survey data showing 76 percent of organizations expect to increase open source AI use. AWS also cites CrowdStrike’s fine-tuned NVIDIA Nemotron model reaching 96 percent valid query accuracy and Gartner’s estimate that agentic workflows multiply token consumption 5 to 30 times.

For builders with data residency rules or cost-sensitive workloads, running agents this way removes third-party API data transfer, single-provider lock-in, and per-seat billing, and charges only for tokens consumed. Model choice becomes a parameter change rather than a platform migration.

AWS points to production use at Ethara.AI, which runs multi-agent orchestration on this architecture for AI engineering and research workflows. Teams evaluating the stack should watch how multi-model routing across Kimi K3, GPT-OSS 120B, and NVIDIA Nemotron 3 Super performs on their own repositories and cost profiles.

What matters

• AWS published guidance for running OpenCode, a terminal-native coding agent, on Bedrock.
• Teams keep data in their own AWS account, avoid per-seat fees, and pay per token.
• Multi-model workflows that route tasks among Kimi K3, GPT-OSS 120B, and Nemotron 3 Super are next.

Why it matters

Multi-model workflows that route tasks among Kimi K3, GPT-OSS 120B, and Nemotron 3 Super are next.

This GenAI News article was prepared in original wording using reporting and materials published by AWS Machine Learning Blog. Source reference: https://aws.amazon.com/blogs/machine-learning/use-open-weight-models-as-your-ai-coding-agent-with-amazon-bedrock/.

Drafted by the GenAI News review pipeline.