Strands Decider 2B
Experimental release from AWS Strands Labs (LoRA adapter on Qwen/Qwen3.5-2B-Base). Uses: routing, tool selection, guardrails, triage. Weaker on long multi-step documents (model card). Open-source alternative to TypeSafe's hosted Jev.
- Input
- text
- Output
- text
- License
- apache-2.0
- Pricing
- input: $0 · output: $0 (free open weights (self-hosted)) source
- Verified
- 2026-10-01
How to call it
| Provider | Model id | Endpoint / URL | Docs |
|---|---|---|---|
| Hugging Face | — | huggingface.co/StrandsAgents/strands-decider-2B-hobson-v19 | — |
Notable capabilities (1)
- Decision model (option scoring): Replaces the LM head of Qwen3.5-2B-Base with a readout/pointer head that scores the supplied options and returns calibrated confidences in one pass: JevBench public 0.723, ~106 ms median latency on an RTX 3090 (VentureBeat). source
Open-weights "decision model" from AWS Strands Labs. It picks among given options or rates on a scale with calibrated probabilities; it does not generate free text. Download from Hugging Face (Apache 2.0).
Timeline entry
- TypeSafe AI releases Jev, a 'System One' decision model that returns typed probabilities instead of text ★★★
On Sept 15, 2026 TypeSafe AI, founded by ex-OpenAI researcher Diogo Almeida, released Jev, which it calls the first "System One model": it takes text or JSON plus typed questions and returns only structured values (yes/no probabilities, choice distributions, scores), in 70–500 ms at $0.042 per…