AWS open-sources Strands Decider, a 2B model that picks an answer and says how sure it is
Strands Labs at AWS has released Strands Decider, an open decision model under the Apache 2.0 licence. It takes Qwen3.5-2B-Base, removes the part that generates text, and adds a small pointer head of about a million parameters that scores the given options in one pass. The model has 1.9 billion parameters, answers in a median 115 ms on an RTX 3090, and also runs on Apple-silicon Macs and on CPU.
It answers three kinds of question: yes or no, pick one of several options, and rate on a scale. The team reports 76.2% accuracy (176 of 231 tasks) on the public set of JevBench, a third-party benchmark for decision models. It also says that on short tasks the model has not seen, answers given with a confidence of 0.9 or more are right about 95% of the time.
A small model that runs on a laptop and reports its own confidence is a cheap first gate for agent steps such as routing, tool choice and guardrails. The team's own guidance is the part to copy: below 0.9, confirm the answer or ask a person.
Drafted with AI from the original source, then checked and edited by Sridhar Mukkandi. Spotted a mistake? Write to hello@tarkika.com and we'll correct it here.