ARC-1: 1.7B local decision model for typed choices, scores and yes/no judgments
A 1.7B-parameter decision model (Qwen3-1.7B-Base + LoRA) that answers choice/score/yes-no questions with calibrated probabilities in ~16-25 ms on a single RTX 4060 Ti, free and fully offline.
Why it matters
A fully-local, sub-20ms open decision model trained on a consumer GPU in 11 days, with honest self-reported numbers where it loses — the first credible open alternative for offline typed judgments and agent guardrails.
What you could build with it
An indie developer building a support-ticket triage bot could run ARC-1 on a single gaming GPU to route tickets to the right team and flag urgent or angry ones with calibrated yes/no probabilities, cutting judgment latency to milliseconds with no API bills.
Does it hold up?
Too early to judge — zero stars and zero HF downloads at publication; the benchmark claims (JevBench 68.4%, 16 ms on a 4060 Ti) are self-reported by the author and await independent reproduction. Weights are CC BY-NC 4.0 (code Apache 2.0).
Built with ARC-1
- ARC-1 quickstart notebook (runs on free Colab GPU)colab · The author's runnable notebook demonstrating ticket-routing, urgency scoring and yes/no judgments on a free GPU.
Learn more
First spotted on github: 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 0.71Mistral 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 0.69DeepSeek 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 0.69kev: 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.