"AWS just released Strands Decider 2B, an open source " ...
Quick Answer
AWS has launched Strands Decider 2B, an open-source decision model that utilizes a modified Qwen3.5-2B-Base architecture.
Quick Take
It achieves a JevBench score of 0.723 with a median response time of 115 ms on an RTX 3090, excelling in short tasks but struggling with long multi-step documents.
Key Points
- Strands Decider 2B has 1.9B parameters and runs on various hardware.
- It provides yes/no probabilities and scores with calibrated confidence.
- Achieves 95% accuracy on unseen short tasks with 0.9+ confidence.
- Apache-2.0 licensed, with full training data and recipe available.
- Designed for model routing and hybrid agent applications.
📖 Reader Mode
~1 min read
AWS just released Strands Decider 2B, an open source "decision model" that never writes a single word. It reads a state plus typed questions and returns a choice, a yes/no probability, or a score, each with a calibrated confidence. The core trick: take Qwen3.5-2B-Base, cut off the language-modelling head, and bolt on a ~1M-parameter pointer head. It scores each option's hidden state against the <answer> position. One forward pass, no decoding loop. The result is fast and honest. v19 hits 0.723 on the JevBench public set (167 of 231) with an ECE of 0.052, and answers in a 115 ms median on an RTX 3090. On unseen short tasks, answers at 0.9+ confidence were right about 95% of the time. The trade-off: long multi-step documents remain its weak spot (hard tier 0.505 vs easy tier 1.000). Apache-2.0, 1.9B parameters, runs on CPU, GPU or an Apple silicon Mac, with the full training recipe and data list public. Built for model routing, tool selection, guardrails and hybrid agents where an LLM makes the hard calls and the decider handles the rote ones...... Full analysis: marktechpost.com/2026/10/01/aws… Model: huggingface.co/StrandsAgents/… Repo: github.com/strands-labs/s… Blog: strandsagents.com/blog/introduci… @awscloud @AWS
— Originally published at x.com
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from WebSearch (Tavily)
See more →全球AI芯片峰会，9月上海见！
The 2026 Global AI Chip Summit will take place in Shanghai on September 22-23, focusing on the evolving AI chip landscape, including the shift from training to inference, the rise of diverse chip technologies, and the restructuring of industry competition. Notable speakers include experts from leading universities and companies, discussing advancements in AI chip architecture and applications.