Amazon Released Open-Source AI Model Strands Decider 2B

The new model provides local agent decision-making capabilities without generating text.

Updated on Oct. 1, 2026 in Artificial Intelligence

Isometric editorial illustration of a sculptural silicon wafer component, representing a technical AI decision-making architecture. — Amazon has released the Strands Decider 2B, an open-source AI model designed for local agent routing and classification tasks without prose generation. AI Illustration. Upload story photo >

Live Poll

Do you believe the release of new open-source AI tools significantly disrupts the current market leaders?

Amazon has released its new Strands Decider 2B open-source AI model, developed by the Amazon Web Services Strands Labs team. The tool is designed specifically for fast local agent tasks like routing and classification.

Why it matters

The model enables rapid, non-prose-based decision-making for local agents, potentially streamlining automated processes. By providing these capabilities as open-source, Amazon introduces a new alternative for developers.

Strands Decider 2B is built on the Qwen3.5-2B-Base architecture and maintains a latency of under 150 milliseconds. The model is optimized for routing and classification tasks and does not generate prose.

The players

Amazon

This is a global technology company that provides cloud computing services and develops artificial intelligence tools.

Amazon Web Services Strands Labs

This is the specialized research and development division within Amazon responsible for building the new AI model.

Hugging Face

This is a prominent open-source community and platform for hosting machine learning models and datasets.

Anthropic

This is an artificial intelligence research organization that competes in the broader AI software market.

The details

The model codebase and training scripts are now accessible through the Hugging Face repository. It is intended to power efficient local agent decisions, moving away from the text-generation focus of many existing AI platforms.

October 2026 marks the target timeframe for AI model leadership goals.

The Tech Race

This release follows the architecture established by the Qwen3.5-2B-Base model. It represents a strategic shift toward specialized, non-prose AI agents rather than general-purpose text generators.

Developers and companies can now utilize this tool to build faster, more efficient automated routing systems. Users may experience improved performance in localized AI tasks due to the reduction in model latency.

The takeaway

The move to release specialized, non-prose models reflects a broader industry trend toward high-speed, functional AI agents. Organizations should consider how such specialized tools might replace or augment existing general-purpose models in their workflows.