Sakhanda Wire
NVDA $153.64 -1.60% MSFT $506.98 -2.13% GOOGL $279.95 -1.33% META $620.09 -1.01% AMZN $256.57 -0.46%
← Back to the news

AWS Releases Open-Source Strands Decider 2B Model

AWS Strands Labs has released Strands Decider 2B, an open-source model designed to make bounded decisions rather than generate text. Based on Qwen3.5-2B-Base, it has 1.9 billion parameters and can select one of several options, estimate a yes/no probability from 0 to 1, or assign a level on an ordered scale. Each response includes calibrated confidence.

The model replaces the base model’s language-model head with a pointer head of about 1 million parameters. It compares representations for the requested options in a single forward pass, avoiding a text-decoding loop. The model can run locally on a CPU, consumer GPU or Apple silicon Mac. Weights, code, a data list and the training recipe are available under Apache-2.0; a pip package provides a command-line interface and HTTP server. The bundled server listens on 127.0.0.1 without authentication, so deployments require an external authentication layer. No hosted inference provider currently offers it.

Version v19 recorded 0.723 accuracy on the public JevBench set, with a 0.342 Brier score and 0.052 expected calibration error. Median latency was 115 ms on an RTX 3090 and 153 ms on an M3 Pro. The developers say it is intended for routing, tool selection, triage and guardrails, not coding, chat or summarization.

Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.

Originally published by MarkTechPost on

Read the original on MarkTechPost ↗

The reporting and the image belong to MarkTechPost and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news