The companies, research groups and people publishing System One models.
100 labs111 models6 verified
Ships Laya, an encoder-based decision model trained with its own RLCD recipe, and its multilingual and typed-decision checkpoints.
@onepass
Brazilian lab building the Julia family of small decision models, starting with the 144M-parameter Julia 1.
SAGEA is a frontier AI lab from Nepal building intelligent systems for the real world. We research and engineer foundation models, reasoning systems, and production AI infrastructure.
nandakishorm.comunclaimed
Nandakishor M (GitHub NandhaKishorM, Hugging Face nandakishorm), CEO of Convai Innovations, which publishes Laya. Vega and its vegaml library are personal releases from his own accounts.
fastino.aiunclaimed
Makers of the GLiNER2 encoder family, extended into open-weight decision models with GLiNER2.5-Decide.
github.comunclaimed
Mark Marosi's Decider family of open System One models. Unofficial listing, compiled from his public releases; the source is github.com/Mapika/decider and huggingface.co/Mapika.
Nokia's applied research team in Sunnyvale builds production agents; its researchers published AnyJev, a training-free layer that turns open LLMs into calibrated decision models (github.com/nokia-applied-research).
autotrust.aiunclaimed
Applied research lab that adds small detachable expert blocks to frozen open models; publishes the JEV and GEV System 1 + System 2 models.
Research group behind CLM, which scores a state against candidate actions with a contrastive objective instead of generating text.
github.comunclaimed
Text-mining group at Harbin Institute of Technology, Shenzhen, known for KaLM-Embedding; publishes JevEmbed and KaLM-Jev, which turn embedding and reranker models into Choice, Score and Noul deciders.
github.comunclaimed
Distils Jev-style typed decisions into Qwen3.5 checkpoints as the JevK5 family.
startlux.comunclaimed
Shanghai company focused on local AI; publishes the StartLux-Decision models in six sizes under a non-commercial licence.
tokenrhythm.aiunclaimed
Beijing company working on model routing and agent post-training; publishes NeoHorse-Jev-4B on its own NeoHorse-1-4B base.
github.comunclaimed
Publishes open-jev, a Jev-shaped DeBERTa decision model trained on public gold labels, with its full training code.
Data-curation company that released Nimble, an open Jev-style LoRA on Qwen3.5-9B, with its data, recipe and benchmark suite.
Bangalore payments company that post-trains and publishes Xor, a family of open decision models served through a TypeSafe-compatible API.
github.comunclaimed
Open-source initiative from Türkiye working towards Turkish language models (MoganBERT-TR), by Furkan Yılmaz, Habibe Aleyna Taşdemir and Muhammed Faruk Gözay. Publishes non-commercial decision models.
Thai AI company behind OpenThai-SystemOne, a Thai-and-English decision model with a 256-way slot head.
Team behind ORG-2, an open-source recorder for AI coding-agent sessions. Publishes the open Wald-4B decision model, first released by co-founder Harry He.
rizzoaiacademy.comunclaimed
Simone Rizzo's online AI academy, run by his company Inferentia SL in Spain; publishes Rizzo Flow, an open, local take on Jev.
jaredpalmer.comunclaimed
Trains and publishes the Kev family, LoRA adapters with a pointer head on frozen Qwen bases that serve the System One API.
github.comunclaimed
Publishes Hopper, a Qwen3.5-4B decision adapter built and measured for JevBench, and Hopper 12B, an adapter on Gemma 4 12B.
Computer-use-agent infrastructure project publishing small Cua-S1 decision checkpoints for GUI actions.
superagent.shunclaimed
Superagent Technologies Inc., a Y Combinator-backed AI security company that publishes the Security-One decision model.
Software engineer in Chennai who publishes OpenJevX, a Laya fine-tune for local yes/no, pick-one and rating decisions, with a Go server and the jevx CLI.
frontiersmind.aiunclaimed
Open-foundation-model company that pre-trains its own Lumma bases and publishes the Lumma-Fev decision models.
sciev.orgunclaimed
Personal research project of Angel Luis Robles-Fernández, a postdoctoral researcher at the University of Kansas; not a company.
Swedish inference provider on EU infrastructure that hosts System One models and published bev, a decision adapter for gating agent commands.
huggingface.counclaimed
An independent developer who publishes Nativ, a Hebrew decision encoder, and its nativ-bench test set.
selfjev.devunclaimed
Julien Wuthrich, a machine learning engineer, publishes SelfJev as an independent personal project. It is not affiliated with TypeSafe.
heman10x-ngu.github.iounclaimed
Personal release by Hemant Kumar (heman10x), an applied-AI engineer in India. Verdict is a small encoder decision model with code and a WebGPU demo.
mertkayacs.comunclaimed
Senior AI engineer who publishes JevAlt, open Jev-API decision models for English, Turkish and German, under Eschatia Labs, the open-source lab label he founded.
xyspace.devunclaimed
XY Space Inc., a Toronto AI implementation firm. yev was built by co-founder Cho Yin Yong, and its weights are published from his personal Hugging Face account.
San Francisco maker of the Foxl desktop agent. Bobcat's weights are published by Sanghwa Na on his personal Hugging Face account.
teximal.comunclaimed
Makes Fort, small open decision models that pick one of your options and put a probability on the call. Its Hugging Face organisation lists one member, Kayode Ojo.
huggingface.counclaimed
An independent ML engineer in Paris who publishes French language and speech models (Vigogne, French Whisper) under his personal Hugging Face account.
posthog.comunclaimed
Product-analytics company. Its Jeeves decision model was built by Nicholas P. Waltz and released under PostHog's verified GitHub and Hugging Face organisations.
empiriolabs.aiunclaimed
Seattle inference and hosting company (EmpirioLabs.ai LLC). Aplomb 1 is its first own model, a multimodal decision model.
Enterprise AI company behind h2oGPTe and the H2O-Danube small language models. Released H2O-Lightning-4B, an open-weight decision model.
Agent and MCP cloud platform. Publishes FlyMyJev typed-decision LoRA packages on Hugging Face.
together.aiunclaimed
Inference platform that published Tev1, a Jev-style Qwen3.5-4B fine-tune, together with the training recipe behind it.
doccy.com.auunclaimed
Doccy Pty Ltd, an Australian telehealth company, publishing on Hugging Face as Doccy Health. Its Solomon model is a general document-decision adapter.
github.comunclaimed
The handle of an individual developer in Tokyo (GitHub hiroki-abe-58) who builds Japanese generative-AI tools that run locally. Not a company.
eulogik.comunclaimed
Edge-AI company founded in 2007, with engineering in Bhopal, India and an office in Dubai. Publishes small open models such as pico-type, Bharat-Tiny-LLM and TinyDoc-VLM.
github.comunclaimed
Australian lab working on trust and verification for LLMs and agents; publishes Cygnet, a typed-decision readout over frozen Gemma-4-12B-it.
browser.lexmount.comunclaimed
Lexmount builds cloud browser infrastructure for AI agents (Lexmount Browser); operated by Beijing Zhuyue Technology Co., Ltd.
Decentralised-AI company that publishes the this-that typed-decision models, fine-tuned from Mapika's decider-2b.
huggingface.counclaimed
Personal account of Emre Vergili, who releases the AUBIN models under the project name Norovox. We found no company behind it.
huggingface.counclaimed
Published Jev-Omni, a multimodal decision classifier over text, image, audio and video on Gemma 4.
github.comunclaimed
Independent researcher (Camellia86 on Hugging Face and GitHub) who published Canopy-Jev, a shared-state decision adapter for Qwen3.8-27B.
@corners-aiunclaimed
Hugging Face organisation of Minyoung Yu, author of the oh-my-jev toolkit, linked to corners.co.kr. Publishes CoCo-Decision, a Korean and English decision adapter.
uprelic.comunclaimed
Berlin company (Uprelic GmbH) behind the Uprelic co-work chat. Strom is the decision model it trained for that product and opened as an API.
staymellow.aiunclaimed
Independent AI engineering studio in Dallas, founded in 2025 by Cole McIntosh.
tokimoa.jpunclaimed
Japanese AI consultancy run by Takuya Omi that publishes the uratori family of Japanese decision models.
jebadiah.aiunclaimed
Texas-based maker of the AINode local AI platform (Jason Brashear); publishes the Jebadiah open decision models in 4B, 9B and 27B.
github.comunclaimed
Individual developer (Hugging Face yunqu, GitHub cloud-qu) who publishes the mentat-sys1 decision adapter; serving code lives in the mentat-asi GitHub organisation.
simplejev.orgunclaimed
A research group (Tianxin Wei, Zhining Liu, Huanxin Sheng) publishing JevAny decision adapters. Unrelated to Featherless AI's simple-jev, despite the similar name.
perplexity.aiunclaimed
AI search company. Releases the pplx-decider decision models as open weights and serves them through its Decisions API.
github.comunclaimed
Publishes Winnow, Gemma 4 fine-tunes for typed decisions that run locally through llama.cpp.
zoowork.aiunclaimed
ZooWork, from Serendipity One Inc., publishes Instinct, a family of single-pass decision models with open 4B weights and a hosted API.
kortexa.aiunclaimed
US multimodal AI studio (kortexa.ai) that released Shingi 27B, a local ternary decision model that also reads images.
oraculum.shunclaimed
Oraculum, LLC calls itself an independent AI research lab. Manchego is published from the oraculumai Hugging Face account of Nicolas W. Schlaepfer.
github.comunclaimed
An independent developer in Belgrade (GitHub mihail-gribov, Hugging Face mihailgribov) who publishes typecastlm.
github.comunclaimed
A pseudonymous individual account (agentmessier on Hugging Face, agentmessier-ai on GitHub); no person or company is named.
hannolabs.aiunclaimed
Builds AI for legal, compliance and investment work (Clause Logic Inc.); publishes the Bosun calibrated typed-decision models and the DecisionBench suite.
loopai.comunclaimed
Independent project behind the OpenJev decision model. Its Hugging Face organisation is run by Abhishek Gahlot and points commercial licensing to Loop AI.
deskmind.devunclaimed
Open-source computer-use project for macOS (Eyes, Brain, Hands, Bench), led by 宗羽 (GitHub gxcsoccer). No company is named.
huggingface.counclaimed
Open research project publishing the Valen multimodal decision models; the code lives on the GitHub account Liuziyu77 (Ziyu Liu), and no company is named as the maker.
github.comunclaimed
An independent developer's personal project (GitHub nagisanzenin, Hugging Face nagisanzeninz) that publishes the Nagi decision models and SDK.
Tokyo-based AI startup (Shisa Inc.) that builds Japanese-first open-weight language models and translation, speech and LLM APIs; Shisa DE-1 is its decision model.
blackdrome.techunclaimed
Small AI research lab founded by Atul Saxena and Divyanshi Sharma, working on agent infrastructure, RL environments and evaluation.
apusai.comunclaimed
The AI lab of APUS, a Chinese technology company based in Beijing and Zhengzhou. Earlier open releases include Linly-70B; publishes the APUS-OpenJev decision models.
github.comunclaimed
An individual developer (Hugging Face Tamkimd, GitHub tamkimd) who publishes the TAMEV encoder decision models; no real name is published.
huggingface.counclaimed
Canadian independent research group (GitHub Gestalt-Lab) working on open reasoning systems; publishes Ornstein fine-tunes and the Jeff decision adapter, written by Dan Lougen.
xnu.appunclaimed
An independent developer (GitHub and Hugging Face everettjf) who publishes ezjev, a Jev-style decision model.
github.comunclaimed
Japan-based GitHub organisation that publishes jqv, an open decision engine on Qwen3 by Hajime Imura. We found no website for it.
telepatia.aiunclaimed
Hugging Face organisation linked to Telepatía, a clinical-AI company in Medellín, Colombia. Publishes Portuguese and Spanish fine-tunes of Laya and Hertz-1, an audio front end for typed decisions.
The InternLM team at Shanghai AI Laboratory; publishes Intern-Decision, multimodal typed-decision fine-tunes of Qwen3.5 in three sizes.
metask.aiunclaimed
Publishes Metask-Jev-4B, a multilingual, calibrated decision model trained with the Nimble objective (Hugging Face account wayfind).
github.comunclaimed
Personal open-weights project by Manjunath Janardhan (manjunathshiva) - the OpenDecider family of calibrated decision models.
jeffhub.aiunclaimed
Independent developer of Jeff, a small open decision model with per-task LoRA adapters (GitHub firelex, Hugging Face mstrasser). A personal project, not a company.
falcons.aiunclaimed
Delaware company (FalconsAI, Inc.) behind the Falconsai models on Hugging Face (content moderation, NLP) and the Model Surgeon tool. Publishes the FalconDec family of small encoder decision models.
agentbull.comunclaimed
Research startup (AgentBull Pte Ltd) building AI agents for decision-makers. Publishes the open-weight Bongard judgment models.
huggingface.counclaimed
An individual's Hugging Face account (profile name "Reflex Labs"; no company, person or contact is published) that releases the vera decision models.
itsnav.comunclaimed
Naveenraj Kamalakannan's one-person Hugging Face organisation (he is a graduate student at NYU). Publishes Plumbify, decision heads built into open LLMs.
@nomeda-labunclaimed
Small Arabic NLP lab (Ahmed Badr, Abdallah Basem Zain, Belal Fathy) publishing the NOMEDA-Decision line.
github.comunclaimed
An individual developer who publishes the OEV decision model from his personal GitHub and Hugging Face accounts.
meragpt.comunclaimed
API platform for small single-task models; serves Decider 1, a hosted decision model that speaks the System One schema.
SynACK is the name of Benjamin Babeshkin's independent AI-agent project. It is not Synack, Inc.
Individual developer (Hugging Face CountingSheep, GitHub Xiaooolong) who publishes Vev, a Jev-like decision model that also reads images.
nomadamas.orgunclaimed
Korean open-source AI hacker house backed by Markr AI. KoJev was released by Jeffrey Dongkyu Kim.
codepawl.comunclaimed
Independent software studio in Vietnam whose research lab, Pawlab, publishes the Tacet typed-decision encoder.
Strands Labs, the experimental arm of Strands Agents, the open-source agent SDK published by Amazon Web Services. Publishes Strands Decider.
ifm.aiunclaimed
Research institute behind the open K2 model families (K2, K2-Think, K2-V2, K2-Horizon). Fine-tuned its smallest K2-Horizon backbone into the decision model K2-Type.
huggingface.counclaimed
Personal Hugging Face account (an individual, not a company) that publishes pjev-2b, a small letter-logit decision model on MiniCPM5-2B.
omnijev.github.iounclaimed
Open research project on Jev-style decision models, run by Li Bobo (GitHub unikcc, Hugging Face BradNLP). Not a company.
huggingface.counclaimed
Dimitris Papadopoulos (lighteternal on Hugging Face) publishes biomedical and Greek-language NLP models; BioDecision is a personal release.
github.comunclaimed
ML scientist who built bev-decider, a 0.4B choice-order-invariant decision model, as a personal project.
jevstyle.comunclaimed
Independent developer (chaoliangUNSW on Hugging Face) who publishes the Jev-Style decision models and the jev-style local server.
Building System One models? Sign up and create your organisation, or claim the page we started for you.