Strands Decider 2B
AWS's open-source 2B-parameter decision model that picks among predefined options with calibrated confidence scores, returning answers in ~100ms on local hardware.
Why it matters
Amazon's Strands Labs entry into the decision-model wave: a Jev-style model that is fully open (weights, training data, scripts) under Apache 2.0 and small enough to run on a local CPU/GPU.
What you could build with it
A DevOps startup could embed the model as a cheap gate inside deployment agents, letting it decide 'safe to auto-merge or escalate to a human' on every pull request, and cut LLM inference spend on routine approvals to near zero.
Does it hold up?
Too early to judge — released Oct 1; AWS reports ~115ms median decisions on an RTX 3090 and 3rd of 33 on JevBench in its size class, but no independent reproductions exist yet. A community GGUF build appearing within 24 hours is an early positive signal.
Built with Strands Decider 2B
- VentureBeat: Amazon unveils a free, fast, open source Jev killerarticle · Coverage of the release with the technical benchmark charts AWS shared ahead of publication, including latency and accuracy comparisons.
- The AI Economy: Why AWS gave Strands a model that can't write a wordarticle · Independent analysis evaluating the model on accuracy (Brier score) and latency, comparing it against TypeSafe's Jev, LocalJev and OpenJev.
- decider-2b-GGUF: llama.cpp-ready GGUF conversionhf · Third-party GGUF conversion of the Decider 2B weights for running in llama.cpp, published within a day of the release.
Learn more
First spotted on article: source.
More AI models
ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71DeepSeek V4-Flash official API: public beta with upgraded agent capabilities, Responses API and Codex supportDeepSeek's official V4-Flash is now a public beta API with massively upgraded agent capabilities - benchmark scores…models · JEV 0.69Mistral Forge: enterprise platform for training and continuously improving proprietary modelsMistral AI launched Forge (Sep 28) — a full-lifecycle model training platform (pre-training, SFT, DPO/ODPO, RL…models · JEV 0.69kev: open, trainable Jev-like family of small decision models on Qwen3.5/3.8A family of small decision models (0.8B to 27B) built on Qwen3.5/3.8 that reproduces Jev's prefill-only typed-decision…models · JEV 0.68GPT-Synopsys: OpenAI and Synopsys multi-year deal to build an AI model that operates EDA toolsOpenAI and Synopsys signed a multi-year agreement on Sep 30 to jointly develop GPT-Synopsys, a specialized model that…models · JEV 0.66
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.