Jev alternatives: hosted models and open weights

Jev alternatives include hosted models and open weights, plus community projects and tools. The registry lists 138 alternatives alongside Jev, for 139 entries.

The registry supplies each summary and Price fact. Capability descriptions and prices are vendor-reported; model pages link labelled evidence. Community projects and tools have their own registry groups, separate from open weights.

Jev

• Jev by TypeSafe AI: A hosted API: send your app's state plus typed questions (Choice, Score or Noul) and get a probability back for each answer. Price (vendor-reported): $0.042 per million input tokens; output free. Source: Announcement → Source: Jev 1.14 → Source: Docs → Source: On OpenRouter → Source: Experimental Box primitive → Compare prices

Hosted

• Decisions API by OpenAI: OpenAI's Decisions API uses GPT-6 Luna to return typed answers for classification, routing and agent actions. Price (vendor-reported): $0.10 per million uncached input tokens; output free. Source: Decisions guide → Source: API changelog → Compare prices
• d1 by Liquid AI: Liquid AI's decision model family. Price (vendor-reported): OpenRouter lists d1 and d1.1 at $0.02 and d1-ultra at $0.042 per million input tokens, output free (as of 28 Oct 2026). Source: Docs → Source: Announcement → Source: Decision Index → Source: Open d1 release → Source: d1.1 on OpenRouter → Source: d1-2B → Compare prices
• Mercury Decide by Inception: Inception's decision model, served as a System One endpoint: send state and typed questions, get a choice, a score or a yes/no answer, each with a probability. Price (vendor-reported): $0.02 per million input tokens at a 50% discount, output free; a free tier remains (OpenRouter, as of 22 Oct 2026). Source: OpenRouter → Source: Announcement → Compare prices
• Solar Decide by Upstage: Upstage's decision endpoint, available through OpenRouter. Price (vendor-reported): OpenRouter lists $0.05 per million input tokens, output $0 (as of 2 Oct 2026); the launch post described a limited-time 50% discount. Source: OpenRouter → Source: OpenRouter announcement → Source: Solar Decide Flash → Compare prices
• Span-01 by Respan: A one-pass classifier for agent traces. Price (vendor-reported): $0.02 per million input tokens; output free. Source: Announcement → Source: Span-02 → Source: Span-02 weights (Apache-2.0) → Compare prices
• Decider 1 by meraGPT: A hosted decision model served only at /v1/systemone, the request format TypeSafe's SDK speaks. Price (vendor-reported): $0.03 per million input tokens; output $0. Source: meragpt.com → Compare prices
• Microsoft-Decision-1 by Microsoft: Microsoft's decision-scoring model for routing, classification, prioritisation, verification and workflow control. Price (vendor-reported): $0.042 per million input tokens; output free. Source: Announcement → Compare prices
• Clef-omni by Cloudflare: The Clef family on Workers AI answers typed questions in one pass. Price (vendor-reported): $0.150 per million input tokens. Source: Cloudflare changelog → Compare prices
• Snowflake Decision by Snowflake: A decision model you call from SQL. Source: Announcement → Source: Decision Fast → Compare prices
• Celeris-1-decision by Celeris: A closed decision model for text, structured data and images that can give a one-sentence reason alongside each probability. Price (vendor-reported): $0.04 per million input tokens; output free. Source: Announcement → Compare prices
• SubQ Snap by Subquadratic: Subquadratic reports a 384K-token context for text and images. Price (vendor-reported): $0.04 per million input tokens; output free. Source: Announcement → Source: Version 1.1 release → Source: Baseten → Source: API models, status and pricing → Source: Baseten launch, 02 Nov 2026 → Source: Snap 1.2 release and retired IDs → Compare prices
• Classify-1 by ScaleDown: Scores text or a whole document against your own choice, yes/no and ordered-scale questions. Source: Docs → Source: OpenRouter → Compare prices
• Clairic by Cloudflare: A new branch of the family: a System One extraction model. Price (vendor-reported): $0.10 per million input tokens on OpenRouter; output free. Source: Announcement → Source: OpenRouter listing → Source: OpenRouter decisions inventory (34 entries, checked 4 Nov 2026) → Compare prices
• Parallel Decisions API by Parallel: A decision engine that searches the live web before it answers. Price (vendor-reported): $1 per 1,000 calls (as of 25 Oct 2026). Source: Announcement → Source: Docs → Source: Benchmark → Source: Benchmark announcement → Source: OpenRouter launch and promotion → Source: OpenRouter decision inventory → Compare prices
• Hume Audio Decisions API by Hume AI: Hume's hosted model answers questions about speech and sound with probabilities. Price (vendor-reported): $0.47 per million input tokens (audio, questions and context); answers not billed. Hume's billing docs, checked 03 Nov 2026. Source: Announcement → Source: Product and playground → Source: Docs → Source: Billing → Source: Audio preparation → Source: Launch changelog → Source: Python SDK → Compare prices
• Zoom Decisions API by Zoom: Zoom's hosted Decisions API and Decisions API Pro accept text, images or both. Price (vendor-reported): $0.042 per million tokens for Decisions API; $0.10 for Pro, per Zoom's product page checked 03 Nov 2026. Source: Product, playground and pricing → Source: Launch changelog → Compare prices
• LingDT-Decision-1.0 by Ant Digital Technologies (蚂蚁数科): Ant Digital Technologies' System One model for choosing answers to business questions, including native multi-select and decisions ordered by dependencies. Price (vendor-reported): Ant lists ¥0 per million input and output tokens, with an original input rate of ¥0.20 per million tokens. Checked 04 Nov 2026. The rate is for the 4B test endpoint during a free trial; no trial end date was established. Source: DTMaaS model page → Source: Official model data and API example → Source: API guide → Source: Launch report (IT之家) → Source: Jev pricing source → Source: Microsoft-Decision-1 source → Compare prices
• Reflex by Morph: Morph's hosted 27B decision model is based on Qwen3.5. Price (vendor-reported): $0.042 per million uncached input tokens; $0.008 cached; output free. Source: Documentation → Source: OpenRouter listing → Source: Launch post → Compare prices
• Grayson by Finic: Finic's hosted fraud and risk decision model returns probabilities for yes/no, single-choice and multiple-choice questions. Price (vendor-reported): $0.035 per million input tokens for ask and score; creating and calibrating scoring models is free. Source: Vendor and model identity → Source: Typed questions → Source: Pricing → Source: API contract → Compare prices
• Zuse by Bineric: A European hosted decision-model family with 9B and 27B API choices. Price (vendor-reported): $0.12 per million input tokens for 9B; $0.75 for 27B; no output tokens. Source: API, limits and prices → Source: Paper and evaluation limits → Compare prices
• Gemini Decision 1 by Google DeepMind: A multimodal decision API for text, images, audio, video and PDFs. Price (vendor-reported): Google reports $0.10 per million input tokens; output free. Shared input billed once, plus each question's instructions and criteria. Source: Launch and billing → Source: API guide → Source: Model limits → Source: OpenRouter listing → Source: Independent test, code and per-answer records → Compare prices

Open weights

• Kev by Jared Palmer: Kev is Jared Palmer's family of open, Jev-compatible decision models, from 0.8B to 27B parameters. Price (vendor-reported): $0.042 per million input tokens and $0 per million output tokens for jaredpalmer/kev-4b on OpenRouter (VENDOR-REPORTED, 2 Nov 2026). Source: GitHub → Source: Kev 1.0 release → Source: Hugging Face → Source: Kev 4B on OpenRouter → Source: GitHub API: stars → Compare prices
• DecisionTune 1.0 by DecisionTune: A local decision model built on ModernBERT-large: give it a state, a question and options, and it returns a probability for each option in one encoder pass. Source: Hugging Face → Compare prices
• Tev1-4B by Together AI: Together's Jev-inspired experiment: a Qwen3.5-4B fine-tune that reads a state, a question and 2 to 24 lettered options. Source: Hugging Face → Source: GitHub → Source: Announcement → Source: Untuned 4B baseline test → Compare prices
• GLiNER2.5-Decide by Fastino: A 340M-parameter encoder decision model from the team behind GLiNER, fine-tuned from GLiNER2. Source: Announcement → Source: Hugging Face → Compare prices
• Laya by Convai Innovations: A non-autoregressive decision engine on ModernBERT-large (421M), with a 322M multilingual sibling. Source: Site → Source: GitHub → Source: Model card and preprint citation → Compare prices
• Nimble by Bespoke Labs: A 9B decision model, LoRA-tuned from Qwen3.5-9B, with its 2,676 training examples published. Source: GitHub → Source: Ollama → Source: Hugging Face → Source: Nimble README → Compare prices
• Decision 1.0 by vLLM Semantic Router: Open decision models from the vLLM Semantic Router project, published as "Decision 1.0: towards open decision foundation models". Source: Hugging Face → Source: Announcement PR → Source: d3 on Hugging Face → Compare prices
• Decider by Perplexity: Perplexity's decision model reads text and images and returns a probability for every answer. Price (vendor-reported): $0.02 (V1.2) and $0.01 (Flash) per million input tokens on OpenRouter, output free (as of 22 Oct 2026). Source: V1.2 announcement → Source: Hugging Face → Source: V1.1 announcement → Compare prices
• Drex 2.0 by Nace.AI: Drex 2.0 returns probabilities over typed answers for text, with open weights under Nace.AI's Open RAIL licence. Price (vendor-reported): $0.03 per million input tokens for Drex 2.0, Drex Max, Max v1.1 and Max Finance, hosted at Nace. Source: Model table → Source: Hugging Face → Source: Drex Max weights and local inference → Source: Drex Max Finance API guide → Source: OpenRouter decision inventory → Compare prices
• H2O-Lightning-4B by H2O.ai: A 4B open model built on Qwen3.5-4B that runs on unmodified vLLM. Source: Announcement → Source: Hugging Face → Compare prices
• Jebadiah v2.1 by frontier-infra: Open decision models that read option-label probabilities in a single pass with thinking switched off: you get a pick and a confidence, not an essay. Source: Release post → Compare prices
• Lumma-Fev-27B by FrontiersMind: A 27B open decision model that answers in English, Hindi, Tamil and Kannada. Source: Release post → Compare prices
• Granite Switch 4.1 by IBM: IBM's switchable models do chat, decisions and speech-to-text from one endpoint. 4.1 adds a general-purpose Decision adapter that speaks the /v1/systemone interface. Source: Announcement → Compare prices
• TYGER by NineNineSix: A decision model for speech. Source: Hugging Face → Source: Announcement → Source: Voice agent → Compare prices
• Talos by Tractian: A decision model that stands between AI agents and the controllers that run factories. Source: GitHub → Source: Hugging Face → Compare prices
• Mammoth 1 by Baseten (Base Labs): Two open decision models from Base Labs, Baseten's research team: mammoth-1 (27B) and elephant-1 (9B), post-trained from Qwen3.5. Price (vendor-reported): OpenRouter lists mammoth-1 at $0.10, elephant-1 at $0.035 and joey-1 at $0.01 per million input tokens, output free (as of 30 Oct 2026). Source: Announcement → Source: Technical report → Source: mammoth-1 on Hugging Face → Source: elephant-1 on Hugging Face → Source: joey-1 → Compare prices
• Qev 0.8B by labqoat: A tiny open decision model that runs locally, built on Qwen3.5-0.8B and reading text and images. Source: Model page → Source: Hugging Face → Compare prices
• Decido 1 family by decido-ai: Six English document-decision packages share the decido API. Source: Decido-1 card and weights → Source: Fast card and weights → Source: Ultra card and weights → Source: Think card and weights → Source: Max card and weights → Source: XL card and weights → Compare prices
• Arc-decide by Arc53 (DocsGPT): The small decision model inside DocsGPT, used for quick judgements such as whether a message is a prompt injection or whether a chunk answers the question. Source: Docs → Source: Hugging Face → Compare prices
• Valen by Valen Team: Open Qwen3.5-based decision models for text, images and video. Source: Site and demo → Source: Hugging Face → Source: 1.5 card and weights → Source: 1.5 public result records → Compare prices
• Fovea-4B by Sivasuthan: An open decision model for images: give it a photo, document, receipt, chart or screenshot plus a typed question and it returns a probability for every allowed answer, served over Jev's System One API. Source: GitHub → Source: Hugging Face → Compare prices
• homi by egma-ai: An open decision model that listens instead of reading: give it an audio clip and a typed question (how does the speaker feel? does the pitch rise?) and it returns a probability for every option, without a transcript. Source: GitHub → Source: Hugging Face → Source: Show HN → Compare prices
• Nimbus by Crele: A decision model for live audio. Source: GitHub → Source: Hugging Face → Compare prices
• GLiDE by Fastino: Fastino’s decision-model family spans 0.4B, 0.8B, 2B, 9B and 27B releases. Source: Hugging Face → Source: API docs → Source: Decision Index → Source: Family announcement → Source: 0.4B card and weights → Source: 9B card and weights → Compare prices
• UltraBERT Curation by Daniel van Strien: A 380M-parameter decision model for data curation: give it a web page, prompt, maths problem or library record and typed questions (topic, format, educational value, difficulty) and it returns a probability for every option. Source: Hugging Face → Source: Recipe → Compare prices
• KnowLine by PelaAI: A fast-moving family of open decision models from PelaAI, trained in an agent-driven data loop, with English and Chinese in mind: from 0.8B up to 35B-A3B, with new generations posted every few days. Source: Hugging Face → Source: Decision Index → Compare prices
• Ru-Jev by besktech (beskvladimir-create): A Russian-first typed-decision model built on mmBERT-base (307M), with a separate scoring head. Source: Weights and v1.3 model card → Source: GitHub → Source: Release metadata → Compare prices
• Intern-Decision by InternLM: Send a shared state, named questions and optional images to these multimodal models. Source: 4B weights and model card → Source: GitHub → Source: Report → Source: Hugging Face upload metadata → Compare prices
• MiniCPM-OCR Jev by OpenBMB: This 2B specialist scores whether a candidate OCR reading matches a document region. Source: Weights and model card → Source: Code and deployment → Source: JevOCRBench → Source: Hugging Face upload metadata → Compare prices
• hs-decide-1-flash by Vectorize (Hindsight): Vectorize's Hindsight team fine-tuned this 0.8B model from Qwen3.5-0.8B for agent memory. Price (vendor-reported): $0.02 per million input tokens on Hindsight Inference; input only. Source: Launch and evaluation → Source: Weights and model card → Source: Hindsight code → Source: System evaluation → Compare prices
• Smaug-Decisions by Abacus.AI: Abacus.AI's 4B decision model answers a typed question about audio, an image or text with a probability per option in one pass. Source: Weights and model card → Source: 0.7B audio tier → Source: LiteRT build → Source: Maker blog announcement → Source: X announcement: 3 Nov 2026 → Compare prices
• REAX 1.3 by REAX: REAX 1.3 runs two model parts on one GPU and averages their per-option probabilities 50/50. Source: Weights, results and upstream notices → Source: REAX Engine → Source: Announcement → Compare prices
• K-Decision-9B by ONYULTEA: A Korean-specialised 9B decision model based on Qwen3.5-9B, trained with int8 LoRA at rank 32. Source: Weights, licence declaration and results → Source: Serving and evaluation code → Source: Weight file inventory → Compare prices
• Saif 0.8B Lightning by GyrTechnology: A 0.8B decision model fine-tuned from Qwen3.5-0.8B. Source: Current card and weights → Source: Release files → Compare prices
• Horopter beta-3 by Fight Fire with Fire Robotics: An encoder-only decision model family built on Ettin, with 0.4B and 1B sizes. Source: 1B model card and weights → Source: 0.4B model card and weights → Source: Runtime and recipes → Compare prices
• TS-PRGuard by typestrict (Lucas Duarte): A 150M pull-request decision model built on TS-Ettin-150M. Source: Model card and package → Source: Model page and audit → Source: Release history → Compare prices
• wps-decision-1 by Kingsoft Office: A d3-flash fine-tune with typed choice, yes/no and score answers for text, JSON, images and videos. Source: Model card and weights → Source: Release history → Compare prices
• d3-mini-agentcode by stonefish-labs: A pre-release 4B d3-mini fine-tune for reviewing code written or run by an agent. Source: Model card, weights and results → Source: Release history → Compare prices
• Lory by eltokh7: A public-preview family for simulated survey respondents, fine-tuned from Kev at 4B and 27B. Source: 27B card and merged weights → Source: 4B card and adapter → Compare prices
• Glimmer System One by Aryan Srivastava: A visual-inspection decision model built around Muse Glimmer 30B. Source: Heads and model card → Source: Code and run records → Compare prices
• laya-slotjudge-m46 by Christophe Cruz, Hussam Ghanem and Hocine Cherifi: A multilingual Laya fine-tune that chooses a proposed value for a dialogue slot or answers not specified. Source: Model card and weights → Source: Code → Compare prices
• Cicero by Butler Compute: A Qwen3.5-35B-A3B decision fine-tune with 3B active parameters. Source: BF16 card and weights → Source: NVFP4 package → Source: Release history → Compare prices
• Adjudex by NewsBang: Qwen3.5 typed-decision models in 9B v5 and 0.8B v5 packages, with a separate 255-option head and noncausal attention in their full-attention layers. Source: 9B v5 card and package → Source: 0.8B v5 card and package → Compare prices
• kai-decision / K-IM-Deiction by Songkai: A Qwen3.5-0.8B model with seven fixed heads for Chinese enterprise group chats. Source: K-IM-Deiction card and merged weights → Source: Package history → Compare prices
• Kev LUX Places by Yale Cultural Heritage AI: Two Kev-4B adapters check whether linked authority records describe the same place and whether a declared parent contains a child place. Source: Equivalence card and adapter → Source: Part-of card and adapter → Compare prices
• d1-qwen-0.6b by Grizzlykw: An independent Qwen3-0.6B text decision starter with a frozen backbone and a trained 1,025-parameter compatibility head. Source: Card and standalone package → Source: Training report → Compare prices
• Jev-style spam LoRA by Abubakar Abid: A Qwen3.5-2B LoRA for one forum-spam question. Source: Card and adapter → Compare prices
• Heimdall D1 V6 by Toby-NoSkillSon: A text-only d1-2B fine-tune that decides whether transcribed speech is for a voice assistant, still unfinished or meant for someone else. Source: Card, weights and MLX runner → Compare prices
• RSI-Jev v7.0-VL family by Shanghua Gao: Qwen3.5-derived decision models with text exits at different depths. Source: Model card and files → Source: 3B package and correction to the 4B release score → Compare prices
• Vica Laya by tory-labs: A Korean-focused Laya multilingual fine-tune. Source: Model card and files → Compare prices
• Spot Snake decision fine-tunes by Lucasgitguy-co: Four local move-ranking fine-tunes for Spot’s Biscuit Snake: d1-2B, d1-3B, MiniMind-JEV and Laya Multilingual. Source: Model card and files → Source: d1-2B card and weights → Source: MiniMind-JEV card and weights → Source: Multilingual card and weights → Compare prices
• JAD-S1-7B-A1B v0.3 by DreamBlooms: A LoRA decision adapter on LLaDA-MoE-7B-A1B-Instruct. Source: Model card and files → Compare prices
• swiftpick-30B-A3B by harishvs: A decision adapter and readout head trained on the text decoder of Nemotron-3-Nano-Omni-30B-A3B. Source: Model card and files → Source: Multimodal runtime and evaluation → Compare prices
• Saif 4B v2 by GyrTechnology: A Qwen3.5-4B full decision fine-tune that returns probabilities for typed questions in one forward pass. Source: Model card and files → Compare prices
• karpo-decision-31B by MachinePulseAI: A text-only Gemma-4-31B-it fine-tune with a linear readout over supplied options. Source: Card and weights → Source: Weight licence and effective date → Source: Separately licensed inference code → Compare prices
• JevStyle 2B by Chaoliang Yan: An independently maintained 2.21B typed-decision model with a frozen vision tower. Source: Card and weights → Compare prices
• Laya spam classifiers by Nikita Kolmogorov: Two multilingual Laya fine-tunes copy Jev judgments for Telegram messages and Farcaster casts or accounts. Source: Telegram card and weights → Source: Farcaster card and weights → Compare prices
• OneJev 31B by OmniJev: A Gemma-4-31B-it decision family served through Flide. Source: Card and weights → Source: 1.6g card and weights → Compare prices
• Copernicus 31B / 57B by Poindexter Labs: Poindexter Labs publishes 31B and 57B structured-decision packages with System One servers. Source: Card, code and weights → Source: 57B card, weights and serving requirements → Compare prices
• Scantron Preview by aidendle94: A research d1-2B fine-tune for proposed agent actions, followed by a 10,240-weight output-head edit. Source: Card, weights and provenance → Compare prices
• Sarvam Prism by Sarvam AI: A Sarvam-105B decision fine-tune, offered as custom FP8 weights and a hosted API. Price (vendor-reported): Sarvam API: ₹4 per million input tokens; output free. Source: Launch and evaluation → Source: Card, runtime and weights → Source: Hosted API limits and price → Compare prices
• Kanana-2-v-decide by Kakao: A text/image decision model built on Kanana-2-v-3b. Source: Model card and weights → Compare prices
• Agastya Decision RAG 2B by Andra Labs (Amartya Sai Andra): A Qwen3.5-2B fine-tune with an Unsloth Clef decision head, intended to choose sections in table-of-contents retrieval. Source: Model card and weights → Compare prices
• termite-decide 0.8B by Tane Piper: A Kev-0.8B fine-tune that judges shell commands for termite. Source: Model card and weights → Compare prices
• Qwen3.5-4B-Router-1M by Maziyar Panahi: A Qwen3.5-4B LoRA and Unsloth decision head for next-action and tool routing. Source: Model card and weights → Compare prices
• Decisio 12B tuned v2 commit by tachara-ai: A Gemma 4 12B fine-tune with merged weights for the decisio runtime. Source: Model card and weights → Compare prices
• Rein r1 family by jmc0: Open 0.8B and 4B decision models that steer a separate chat model using rules supplied per request. Source: Model card and weights → Source: 0.8B card and weights → Compare prices
• Vev 4B v2 by Wang Xiaolong: A Qwen3.5-4B text/image decision fine-tune with merged weights and a Jev-like API. Source: Model card and weights → Compare prices
• Compliance Decisions 0.8B by souvickdascmsa019: A Qwen3.5-0.8B LoRA plus an Unsloth Clef head for fixed sheets of regulatory and general questions. Source: Model card and weights → Compare prices
• JDI-Decision-V1 by JD Industrials: A Qwen3.8-27B decision system that combines vision features with frozen PaddleOCR v5 transcriptions. Source: Model card and weights → Compare prices
• Jeva 4B by Qi Li and Xinchao Wang (NUS): A text decision model with Qwen3.5-4B backbone, adapter and a separate decision head. Source: Model card and weights → Compare prices
• RASD Laya track classifier by Galal Ban Amr: A Laya fine-tune that labels English descriptions of simulated sensor tracks as drone, bird, aircraft or unknown. Source: Model card and weights → Source: Recorded metrics → Source: Synthetic dataset → Compare prices
• MoDe S / M / L by newrexlabs: Three sparse MobileMoE decision fine-tunes for English text or JSON. Source: Small card and weights → Source: Medium card and weights → Source: Large card and weights → Source: Inference code → Source: Licence → Compare prices
• Laya Minecraft by Thomas Tesche: An English Laya fine-tune for a Minecraft agent. Source: Model card and weights → Source: Training and evaluation report → Source: Agent source → Compare prices
• medgemma-decide-4b-v2 by Amaljith: An English medical-text decision fine-tune continued from a MedGemma 1.5 4B derivative. Source: Model card and weights → Source: Per-question results → Source: Derivative notice → Compare prices
• medical-jev-4b by JuneYao: A Qwen3.5-4B LoRA for supplied medical answer choices, mainly in Chinese. Source: Model card and adapter → Source: Source and evaluation → Source: Released runtime record → Compare prices
• LingeeJevMulti-4B-v1 by Ashan0525: A Qwen3.5-4B derivative that reads text or images and returns probabilities for Choice, Noul and Score questions without generating an explanation. Source: Card and release files → Source: Licence → Compare prices
• Logits Sigma-1 / Duo by Logits: Two text decision packages. Source: Sigma-1 card and weights → Source: Duo card and weights → Source: Duo source-data disclosure → Compare prices
• d1-2B RAG Verifier by landtml: A d1-2B LoRA for checking whether a supplied answer is supported by its source. Source: Card, weights and evaluation → Source: Scoring code → Source: Licence → Compare prices
• LocHeal Verdict 2B by kinasr: A d1-2B fine-tune for choosing a replacement DOM element when a browser-test locator breaks. Source: Card and weights → Source: Licence → Compare prices
• Amada by LiamDB: A CPU decision model family. Source: Code, model description and limits → Source: 3.1 release and weights → Source: Licence → Compare prices
• JevAlt by Mert Kaya (Eschatia Labs): Three Intern-Decision-4B fine-tunes: Deem-4B for English, Karar-4B for Turkish and Wähler-4B for German. Source: Code, licence and serving → Source: Deem English card → Source: Karar Turkish card → Source: Wähler German card → Source: Recorded live requests → Compare prices
• Rein-NPC-4B by jmc0: A Qwen3.5-4B chat checkpoint with a decision readout trained through 64 added token embeddings. Source: Card and weights → Source: Licence → Compare prices
• Laya Descent v4 by Eren Terakye: A Laya fine-tune that copies a scripted lunar-lander autopilot for Low Gate. Source: Card and checkpoints → Source: Training report → Source: Low Gate benchmark → Compare prices
• Chinese Decision Pointer 4B by quincyqiang / yanqiangmiffy: A Qwen3.5-4B LoRA fine-tune with a pointer head that scores markers beside the supplied candidates. Source: Card and weights → Source: Scores and request counts → Source: Serving and training code → Compare prices
• Aster-JEV-27B by k2021211255: A full-parameter Qwen3.8-27B vision decision fine-tune with a separate head and a frozen evaluation engine. Source: Card, code and weights → Source: Pinned public results → Source: Weight licence → Compare prices
• Jevora 9B by kadirbekovvv / doxanocap: A Qwen3.5-9B LoRA decision model with a separate 255-option head. Source: Current V1.6 card → Source: Pinned package → Source: Loader and licence → Source: Package manifest → Compare prices
• Domyn-Edge-Jev-Decision by Mahesh Deshwal: A 3.93B full fine-tune of Domyn-Edge-SFT for text decisions. Source: Card and weights → Source: Evaluation report → Source: Pending JevBench adapter → Compare prices
• Clashbot d1-ft-joint-tac by Jyler: A fine-tune of Liquid AI’s d1-omni-600M for a Clash Royale course project. Source: Card, weights and code → Source: Licence → Compare prices
• QEV Compact by Ken Jo: Two pruned QEV v3 models for text and image questions. Source: 2B compact card and weights → Source: 0.8B compact card and weights → Source: Licence → Compare prices
• Julia 1 by Supersonic Labs: A 144.3M-parameter multilingual encoder adapted from mmBERT-small to score supplied options. Source: Card, runtime and weights → Source: CPU results → Source: Checkpoint provenance → Compare prices
• Strands Decider 2B v1 by Strands Labs: A LoRA adapter and option-scoring head on Qwen3.5-2B-Base for Choice, Noul and Score questions. Source: Card, adapter and run records → Source: Licence → Source: Runtime and training code → Compare prices
• ZeroPoint by Javier Marín: A CPU decision model that scores supplied options against a fixed zero-energy none state. Source: Code and bundled model → Source: Saved comparison → Source: Licence → Source: Model manifest → Source: Package releases → Compare prices
• MANTA by Gevaert Lab: Medical decision models in two sizes, built on Qwen3.5-4B and Qwen3.8-27B with a trained decision head. Source: 4B card and weights → Source: 27B card and weights → Source: 4B configuration → Source: Licence → Compare prices
• N-Q-26M-JEV by Q Project: A small Russian/English decision model with a custom hybrid backbone and decision head. Source: Card and weights → Source: Licence → Source: Runtime → Compare prices
• GameDecider by Holdout Labs: An English game-scene decision model derived from Ettin-400M, with ONNX files and a Python runtime. Source: Card and files → Source: Training run record → Source: Data sources → Source: Licence → Compare prices
• MNN-Pocket-0.3B-Dec by Taobao MNN: A text decision package with the MNN-Pocket-0.3B backbone, an unmerged LoRA adapter and a pointer head. Source: Card and weights → Source: Runtime → Source: Calibration record → Source: Licence → Compare prices
• ARC 1.0 by Rayline: A Qwen3.5-0.8B-derived router that predicts model success and token use separately from the prices supplied by an application. Source: Card and release files → Source: Policies → Source: Provenance → Source: Licence → Source: Public plugin and limits → Compare prices
• GS1 by BIGAI: A four-size family with Qwen3.5 hybrid backbones and native decision heads. Source: 2B card and code → Source: 0.8B package → Source: 4B package → Source: 9B package → Source: 9B request limits → Source: Licence → Compare prices
• ANT by Semantique Team: Trained candidate-scoring heads for frozen Qwen3.5-0.8B and 4B backbones. Source: ANT-4B card → Source: ANT-08B card → Source: 4B metrics → Source: 08B policy → Source: Licence → Compare prices
• decider-ja 0.5B by ken11: A Japanese decision adapter and pointer head trained on Sarashina2.2-0.5B with the Strands Decider recipe. Source: Card and packages → Source: Training configuration → Source: Request configuration → Source: Licence → Compare prices
• Ainvest Decision 31B v2.0 by Ainvest-World: A text-only Gemma 4 31B fine-tune that scores supplied option tokens without generating an answer. Source: Card and inference package → Source: Licence → Source: Third-party terms → Source: Weight manifest → Compare prices
• newTokenEra decision 27B by NewTokenEra: A Chinese/English post-trained Qwen3.8-27B decision model with full BF16 weights. Source: Card and weights → Source: Licence → Source: Provenance → Source: Evaluation record → Source: Typed inference implementation → Compare prices
• ProbeLoop Laya stated limits by submsk: A provisional Italian sentence classifier fine-tuned from Laya multilingual. Source: Card and weights → Source: Licence → Source: Checksums and provenance → Source: Pilot and integration record → Compare prices
• femto-jev by igorktech: A small typed-decision encoder for text, with a separate femto-asr speech frontend. Source: Card and weights → Source: Inference code → Source: Licence → Source: Configuration → Compare prices
• Laya medical specialty by AI Anytime: A Laya fine-tune that assigns clinical notes to a fixed menu of medical-specialty and document-type labels. Source: Card and weights → Source: Evaluation record → Source: Training and serving toolkit → Compare prices

Tooling

• Ollama v0.35 by Ollama: Ollama added native decision models through /v1/systemone, based on TypeSafe's Jev API, so you can run one on your own machine. Source: Release notes → Compare prices
• prismhq/jev-router by prismhq: A LiteLLM-based router with an optional Jev decision backend: when a TypeSafe API key is configured it can use Jev to pick a model per request, and otherwise falls back to its own rules-based decider. Source: GitHub → Compare prices
• mizorewww/laya-mlx by mizorewww: Runs the open Laya decision models natively on Apple Silicon with MLX: no PyTorch, no cloud API, zero output tokens. Source: GitHub → Compare prices

Community

• SemIf (formerly OpenJev) by TheoLeeCJ: Reproduces Jev's interface pattern with open models: it reads typed option probabilities directly from a model, with no answer sentence and no decoding loop. Source: openjev.com → Source: TheoLeeCJ/SemIf-OpenJev → Compare prices
• bnsd55/jevmlx by bnsd55: Parallel constrained decisions for local models through mlx-lm on Apple Silicon: the schema is a set of typed fields, the context is prefilled once and broadcast across one batch row per field. Source: GitHub → Compare prices
• kw2828/OpenJev by kw2828: A browser and local playground for decision experiments on memory, uncertainty and control, using Qwen-family models and returning option probabilities. Source: GitHub → Compare prices
• akash-kamat/system-one-gemma by akash-kamat: Gemma 3 270M with a scoring head, producing a decision in a single forward pass with no text generation. Source: GitHub → Compare prices
• leofmarciano/groot by leofmarciano: Groot 1 S, a 149M-parameter open typed-decision model that speaks the System One request format and reads states of up to 64k tokens in the maker's tests. Source: GitHub → Compare prices
• jwuthri-cloud/SelfJev by jwuthri-cloud: A self-hosted decision model with a Jev-compatible API: you run the server on your own GPU, and a 230 MB vision adapter on Qwen3.5-4B reads text and images. Source: GitHub → Source: Hugging Face → Compare prices
• SecJev-4B Spot2 V3 by macmacmacmac: A Qwen3.5-4B-Base LoRA and Kev head for authorising outgoing tool calls. Source: Card, licence and records → Compare prices
• CodeSync Knack by DevsHub: A local CodeSync bundle with 29 fixed judgments about user turns and shell commands. Source: Card and bundle → Source: Standalone runner → Compare prices

Decision models versus classifiers · Evaluate on your own task · Local decision models · Compare decision models · Compare all published prices · Compare JevBench scores · What is a decision model?